Firm Growth and Disclosure: An Empirical Analysis

Author(s): Inder K. Khurana, Raynolde Pereira and Xiumin Martin

Source: The Journal of Financial and Quantitative Analysis, Vol. 41, No. 2 (Jun., 2006), pp.
357-380
Published by: Cambridge University Press on behalf of the University of Washington School of
Business Administration
Stable URL: http://www.jstor.org/stable/27647251
Accessed: 25-03-2015 02:03 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact [email protected].

Cambridge University Press and University of Washington School of Business Administration are collaborating with
JSTOR to digitize, preserve and extend access to The Journal of Financial and Quantitative Analysis.

http://www.jstor.org

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 JOURNAL OF FINANCIAL AND QUANTITATIVE ANALYSIS VOL. 41, NO. 2, JUNE 2006
COPYRIGHT 2006, SCHOOL OF BUSINESS ADMINISTRATION,UNIVERSITYOF WASHINGTON, SEATTLE,WA 98195

FirmGrowth and Disclosure: An Empirical

Analysis
Inder K. Khurana, Raynolde Pereira, and Xiumin Martin*

Abstract
Extant research posits that information asymmetry and agency issues affect the
theoretical
cost of external financing and hence impact the ability of firms to finance their growth
opportunities. In contrast, the literature on disclosure policy posits that expanded and
credible disclosure lowers the cost of external financing and improves a firm's ability to
pursue potentially profitable projects. An empirical implication is that disclosure can help
firms grow by relaxing external financing constraints, thereby allowing capital to flow to
positive net present value projects. This paper empirically evaluates this prediction using
firm-level data over an 11-year period. As anticipated by theory, we find a positive relation
between firm disclosure policy and the externally financed growth rate, after controlling
for other influences.

I. Introduction

The corporate finance literature emphasizes the importance of information

asymmetry and agency costs in influencing firm growth through their impact on
the efficiency of firms' investments (Stein (2003)). In particular, these distortions
serve to constrain firms' access to lower cost external financing and hence limit a
firm's ability to pursue potentially profitable projects (Demirguc-Kunt and Mak
simovic (1998)). However, a related literature suggests that disclosure policy is a
a firm can lower its cost of external
curative mechanism through which financing
and improve its ability to fund growth opportunities (Verrecchia (2001), Bush
man and Smith (2001), and Stein (2003)). An implication of this line of inquiry
is that it anticipates a association between a firm's disclosure policy and
positive
its realized growth rate. This paper empirically evaluates this prediction.

Myers and Majluf (1984) argue that information asymmetry serves to in

crease the cost of external and may therefore force firms to forgo poten
financing
tially profitable As a result, they posit that in the presence of information
projects.

*Khurana, [email protected], Pereira, pereirar@u. mi s souri.edu, and Martin, xxz55@mizzou

.edu, School of Accountancy, College of Business, University ofMissouri, Columbia, MO 65211. We
thank Paul Healy (the reviewer), participants at the 2004 Financial Management Association meeting
and the 2005 American Accounting Association midwest region meeting, as well as seminar partici
pants at theUniversity ofMissouri-Columbia.
357

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
358 Journal of Financial and Quantitative Analysis

a firm's will be constrained to its internal resources. In a sim

asymmetry, growth
ilar vein, agency conflicts can also influence a firm's realized rate. For ex
growth
ample, Myers (1977) identifiesa "debt overhang" problem where managers may
forgo positive net present value (NPV) projects in the presence of riskydebt.] In
light of distortions resulting from information asymmetry and agency conflicts,
firms may confront a higher cost of external to fund their investment
financing
projects.
Extant that an expanded can
theory also posits and credible disclosure policy
improve investment efficiency by mitigating information asymmetry and agency
conflicts.2 A higher level of disclosure serves to reduce the cost of external financ

ing by reducing information asymmetry between investors and managers. On the

other hand, Bushman and Smith (2001) posit a governance role for disclosure,
arguing that it affords a mechanism for investors to monitor insiders. Specifically,
Bushman and Smith ((2003), p. 68) contend that an expanded disclosure policy
"contributes directly to economic performance by disciplining efficient manage
ment of assets in place (for example, timely abandonment oflosing projects),
better and reduced of
project selection, expropriation investors' wealth by the
managers." Given the role of disclosure in ameliorating information asymmetry
and agency conflicts, extant theory contends that an expanded disclosure policy
should serve to improve a firm's access to lower cost external to fund
financing
its growth
opportunities.3
If disclosure is effective in enhancing firm access to external funds, then we

anticipate disclosure to improve a firm's ability to invest in potentially profitable

for growth.
projects We examine this prediction using a cross section of U.S. firms
over an from 1984-1994. we examine the relation
11-year period Specifically,
between a firm's disclosure and its externally financed rate.
policy growth
To proxy for theoverall level of disclosure policy thata firmadopts, we rely
on financial evaluations of firms' disclosure and use
analysts' practices analyst
rankings of overall firm disclosure as in the Association of Investment
reported
Management and Research's Annual Reviews of Corporate Reporting Practices

(AIMR reports).4 The benefit of using the scores assigned by the analysts is that
a measure
they provide ready off-the-shelf that has been widely used in prior
research as a measure of corporate disclosure Further
comprehensive practices.
more, prior studies (e.g., Yu (2005)) find thisdisclosure metric tobe well behaved
in that it is statistically significant in the predicted direction with the dependent
variable of interest. However, we recognize that a firm's disclosure is not
policy

1
Jensen and Meckling (1976) describe another agency problem whereby managers acting in the
interest of equity holders may extract value from debt holders by investing in riskier projects after
debt is in place. This "asset substitution" problem will serve to further increase the cost of external
financing.
2See Verrecchia (2001) for a review of the theoretical literature on voluntary disclosure.
3Prior empirical studies largely examine the relation between firm disclosure policy and cost of
external capital (Botosan (1997), Sengupta (1998), and Leuz and Verrecchia (2000)). In contrast, this
paper focuses on how disclosure policy contributes to firm growth.
4Empirical research using this database documents considerable variation in the level of disclosure
across firms (Lang and Lundholm (1993)). Lundholm and Meyers (2002) cite several studies that
use these disclosure rankings as a measure of a firm's disclosure
policy. For example, Lang and
Lundholm (1996) provide evidence that firms with higher disclosure rankings have less dispersion
among individual analyst forecasts, implying thatmore disclosures reduce information asymmetry.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 359

exogenously determined. Prior research shows that a firm's disclosure policy is

a function of firm-specific characteristics that influence the benefits and costs re
lated to an expanded disclosure policy (Lang and Lundholm (1993)). To address
identification issues, our empirical takes into account the endo
potential analysis
geneity of our disclosure metric.
We follow the approach developed byDemirguc-Kunt andMaksimovic (1998),
(2002) to measure the extent to which a firm's is externally financed.5 An
growth
advantage of using this approach is that it estimates the external financing need

of each individual firm, while at the same time controlling for factors that may
affect the demand for external we calculate for each sample
capital. Specifically,
firm the rate atwhich itcan grow, using i) only its internalfunds or ii) its internal
funds and short-term We then compute the extent to which a firm's
borrowing.
actual rate exceeds each of these two estimated rates and use a simulta
growth
neous to explicitly model excess rates and disclosure as
equation system growth
determined variables. As an additional test,
endogenously dependent sensitivity
we measure disclosures to the measurement of firm growth to avoid the con
prior
cern of simultaneity bias thatmay arise iffirmswith higher externally financed
disclose more.
growth
a sample of 1,436 observations over an 11-year from
Using firm-year period
1984 to 1994, we find that analyst rankings of overall firm disclosure are posi
tively associated with firm growth supported through external financing. These
results after controlling for several firm characteristics linked to external
prevail
needs. In addition, the results are robust to alternative model speci
financing
fications. Overall, we find that disclosure facilitates firm growth. Our finding
is consistent with the notion that disclosure affects firm growth by improving a
firm's access to lower cost externalfinancing.
Overall, our results add to the growing literature on corporate finance and
Recent research the role of institutional factors such as a
growth. emphasizes
environment on investment and For exam
country's legal efficiency growth.6
ple, Levine and Zervos (1998), Rajan and Zingales (1998), and Demirguc-Kunt
and Maksimovic (1998) explore the relation between financial development of
countries, industries, and firms, their on the cost of
respectively, through impact
external financing. However, Stein (2003) notes that investmentdistortions aris
and agency conflicts can impact even in
ing from information asymmetry growth

settings where a country's institutions such as the legal, auditing, and contracting

environment are "highly evolved." In the spirit of this observation and in contrast
to recent studies, this paper focuses on firms in the U.S. where, despite highly

5An alternative
approach is to focus on corporate investment demand along the lines of Fazzari,
Hubbard, and Petersen (1988) and compare the empirical sensitivity of investment to cash flow across
a group of firms sorted according to a proxy for disclosure. Given that extant research (e.g., Almeida
and Campello (2002), Alti (2003), Cleary (1999), Erickson and Whited (2000), Kaplan and Zingales
(1997), and Poterba (1988)) has questioned themeaning of the cash flow sensitivities of investment
on theoretical and empirical grounds, we sidestep these issues by focusing on the impact of disclosure
on firm growth obtained through external financing. As we indicate later, the approach we use in
our study estimates the excess growth made possible by external financing for each firm by directly
identifying firms that cannot internally fund their investment.
6La Porta, Lopez-de-Silanes, Shleifer, and Vishny (2000a) also point to the role of the legal en
vironment as an institutional governance mechanism in limiting wealth expropriation by inside man
agers and owners at the expense of minority shareholders.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
360 Journal of Financial and Quantitative Analysis

and other institutions, information asymmetry and agency con

developed legal
flicts can result in investment distortions. This paper extends Demirguc-Kunt and
Maksimovic's (1998) methodology to address the influenceof disclosure on firm
growth.
and extends prior research in two ways.
Our study also complements First,
research largely focuses on the relation between cost of capital and disclo
prior
sure (Botosan (1997)). However, in the absence of a theoretical pricing model
thatmaps a firm's disclosure level to cost of capital, Lang (1999) contends that
we need to seek additional "confirming evidence using alternative approaches"
before we can draw conclusions the impact of disclosure This
regarding policy.
paper addresses thispoint by linking a firm's growth to itsdisclosure policy. Fur
thermore, it also the measurement associated with a firm's
sidesteps problems
cost of capital. Second, an examination of the relation between cost of capital
and disclosure disregards the role of disclosure in mitigating agency problems.
Recent research by Bushman and Smith (2001) emphasizes thatdisclosure policy
also affects a firm's cash flows the managerial diversion of resources
by limiting
as well as other agency conflict-related Therefore, the focus on firm
problems.
a more measure that captures the theoretical role of a
growth provides complete
firm's disclosure policy.
The remainder of this paper is organized as follows. In Section II, we elab
orate on the role of disclosure and financed and our
externally growth, develop
testable hypothesis. Section III describes our methodology and data. Section IV
presents our empirical findings, and Section V concludes the paper.

II. External Financing, Disclosure, and Growth

Adverse selection costs from information asymmetry can prevent a

resulting
firm from raising external funds to undertake new investments. and Majluf
Myers
(1984) argue that firmswith shortages of cash flow and liquid assets might ac
tually forgo profitable investment spending rather than issue mispriced securities
to fund the investment. these firms may have investment
Consequently, untapped
opportunities thatwould increase firmvalue if sufficientfunds could be gener
ated. Prior research also posits that agency conflicts increase the cost of external

financing. For example, Myers (1977) posits thatmanagers may forgo positive
NPV projects in thepresence of risky debt. Given thispotential conflict, creditors
may price this risk, resulting in a higher cost of borrowing.
Extant theory also posits that an expanded disclosure policy serves to im

prove firms' access to lower cost external financing by ameliorating information

asymmetry and agency conflicts. Verrecchia (1983) points to the role of disclo
sure in adverse selection costs associated with information
limiting asymmetry.
More recent research by Bushman and Smith (2003) highlights how disclosure
can affect the investments, and value-added of firms both low
productivity, by
ering the cost of external financing and
by reducing agency costs. Under this

framework, disclosure not only reduces adverse selection, but also plays a gov
ernance role in improving investors' to monitor firm performance and to
ability
better evaluate managerial performance. The overall is that managers
implication

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 361

are more assets

likely to efficiently manage in place and invest in positive NPV

projects in the presence of a credible and expanded disclosure policy.7

To the extent that a firm's disclosure policy mitigates the problems arising
from information asymmetry and agency conflicts between managers and outside

investors, disclosure improves of firms'

the efficiency investments by lowering
their cost of external Three are noteworthy here. First, despite
financing. points
the benefits of an expanded disclosure policy, extant theory points to the exis
tence of an interior optimal level of disclosure. This result is obtained because
disclosure can be costly to firms (Verrecchia (1983)). For example, if increased
disclosure reveals information to competitors or others who interact strategically
with the firm, itmay cause the firm to lose or
competitive advantage bargaining
power (Admati and Pfleiderer (2000)). In the presence of both benefits and costs
of disclosure, extant theory suggests that firms will
weigh the benefits and costs
of disclosure to reach an disclosure level.8
optimal
indicates that proprietary costs may lead firms to
Second, prior research
avoid disclosure of firm-specific information. For example, extant research
public
depicts privately placed debt as inside debt given the close proximity in the rela
tionbetween theborrower and lender (Bhattacharya and Chiesa (1995), Campbell
(1979), and Rajan (1992)). Viewed as such, extant theoryposits thatborrowers
will turn to private debt markets when the costs associated with publicly reveal
information are substantial. In other words,
ing proprietary (firm-specific) pri
vate debt provides a channel that allows borrowers to benefit from the disclosure

of firm-specific information to private debtholders without suffering the adverse

consequences of public disclosure.9
an expanded access to lower
Third, while disclosure policy may improve firm
cost the resulting benefit may not be equal across all firms. For exam
financing,
ple,
a firm with sufficient internal resources relative to its investment opportunities
is less likely to benefit from an expanded disclosure policy. Alternatively, a firm
with limited internal funds relative to its investment opportunities will benefit
from an expanded disclosure policy ifmore disclosure improves thefirm's ability
to pursue access to lower cost external
potentially profitable projects by providing
financing.
In summary, extant theory establishes that disclosure will affect firm growth
the external financing channel. Therefore, to empirically evaluate this re
through
lation we need to distinguish between a firm's internally and externally financed

7Lang and Lundholm (2000) report a dramatic increase in disclosure activity around seasoned
equity offerings for a sample of 41 firms. However, firms that substantially increase their disclosure
activity in the six months before the offering experience price increases prior to the offering relative to
the control firms, but suffer negative returns relative to the control firms subsequent to the announce
ment, suggesting that the increased disclosure activity for some firmsmay have been hype that enables
them to attain a lower cost of equity capital.
8Consistent with the existence of an optimal level of disclosure, prior empirical studies show
that the level of voluntary disclosure is associated with the costs and benefits of disclosure. For
example, Lang and Lundholm (1993) find that the level of disclosure is positively related to the level
of information asymmetry. Similarly, Bamber and Cheon (1998) find that disclosure levels decline
with proprietary and litigation costs.
9The empirical implication is that the effect of disclosure on externally financed growth may be
related to the type of external financing. In Section IV.D, we conduct additional analysis to test this
implication.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
362 Journal of Financial and Quantitative Analysis

rates. If disclosure supports a rate through the external fi

growth higher growth
channel, then we should observe a
positive association between the exter
nancing
nally financed growth rate and the level of firmdisclosure. The testable hypothesis
can be stated (in the alternate form) as follows.

HI. A firm's externally financed growth rate is positively associated with the level
of disclosure, ceteris paribus.

III. Methodology and Data

A. Empirical Model

Our hypothesis thatexternally financed growth and disclosure are positively

related is based on the theoretical argument that an expanded disclosure policy im

proves firms' access to lower cost external financing. It is plausible that firms that
are growing and performing well may also be more forthcoming in their disclo
sures to firms that are doing poorly. To the extent that these cause and
compared
effect relations are feasible, disclosure and externally financed growth variables
are to be determined. To account for the simultaneity, we estimate
likely jointly
the relation between disclosure and externally financed growth (EFG) using the
following simultaneous equation system,10

= + <52CORRi7+ S3ROAit + ?4STDfY

(1) DISCL,, 6o + SilMVEit
+ ^DISCL/?-i + 66EFGit + uit,

= a + piDISCLit + P2DIVit/TAit + foANlit/NSit

(2) EFGiY
+ ?4ANSit/TAit + ftLOG-TA-,
+ ?Q/i + ?FINft/TA/f + ^,

where DISCL is a measure of a firm's overall level of disclosure; LMVE is the log
ofmarket value of equity (a proxy for size); CORR is thehistorical correlation be
tween annual returns and earnings computed over the preceding 10 years (a proxy
for information asymmetry); ROA is return on assets (a proxy of performance);
STD is standard deviation of annual market-adjusted stock returns over the pre

ceding five years (a proxy for performance variability); EFG is a measure to cap
turegrowthmade possible by external financing; DIV/TA is total dividends/total
assets; NI/NS is earnings after interest and taxes/net sales; NS/TA is net sales/total

assets; LOG_TA is the natural log of total assets; Q is Tobin's Q; and FIN/TA is
equity and debt issuances/total assets.
All variables (except EFG) inmodels (1) and (2) are averaged over the same
time span over which EFG is computed. Model (1) specifies disclosure policy as a
function of firm growth, lagged values of the disclosure score, and four other vari
ables that prior research indicates are related to disclosure. Larger firms tend to
disclose more because of greater demand for information; hence, the coefficient

10We also perform two specification checks for alternative control variables in the disclosure re
gression. First, we include a dummy variable equal to one forwhether a firm issued equity or debt in
the current year, and zero otherwise. Second, we exclude performance variability from our regression
model. The results are not sensitive to any of these variations.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 363

on LMVE is expected to be positive. Lang and Lundholm (1993) find a nega

tive relation between the level of the earnings-returns correlation and the firm's
disclosure level and suggest that disclosures tend to be high when earnings fail
to capture valuation-relevant information. Hence, the coefficient on the variable
CORR is expected to be negative.
Disclosure may be positively related to firm performance in the face of ad
verse selection. That is, firms that exceed a certain profitability thresholdwill
disclose more, while those below the threshold will disclose less. Hence, the co
efficienton the variable ROA is expected to be positive. Similarly, performance
variability may lead to improved disclosure because it increases a firm's vulnera

bility to legal action (Lang and Lundholm (1993)). Hence, the coefficient on the
variable STD is expected to be positive.
Model (2) uses a firm's externally financed growth rate as a dependent vari
able and regresses it on a set of firm characteristics related to both external fi

nancing needs and our test variable proxying for firm-level disclosure. Because
HI predicts a positive relation between disclosure and externally financed growth,
thevariable DISCL is expected tohave a positive sign inour regressionmodel (2).
An advantage of the two-stage least squares (2SLS) estimation is that it yields
consistent parameter estimates because the fitted values of DISCL all ex
using
ogenous variables in equations ( 1) and (2) are uncorrelated with the error term in
model (2).
Demirguc-Kunt andMaksimovic (1998), (2002) point out thatat present we
do not have an explicit theoretical model that links firm characteristics to exter

nally financed growth. In the absence of such a model, the variables in model

(2) attempt to control for factors that can influence externally financed growth,
our variable, in two ways. First, we control for factors that capture
dependent
the availability of internal and external funds. The more the internal funds, the
lower the demand for external sources of funds. Second, we control for factors
that reflect the level of growth opportunities.
In regressionmodel (2), thevariables DIV/TA, Z\NS/NA, Z\NI/NS, LOG_TA,
and FIN/TA control for the extent of availability of internal/externalfunds,while
the variable, Tobin's Q controls for growth opportunities. Firms that pay more
dividends (as a proportion of total assets) (DIV/TA) are viewed to have excess
cash relative to their investment needs (Demirguc-Kunt and Maksimovic (1998)).
In our context, the implication is that externally financed growth is likely to be
smaller for firms with values of DIV/TA; hence, the coefficient on the vari
higher
able DIV/TA is expected to be negative.
Following research, we also control for changes in firm performance,
prior
u
namely, changes in profitmargin and asset turnover. Fairfield and Yohn (2001)
note that "profit margin measures firms' to control the costs incurred to
ability
generate revenues" 372). As such, an increase in profit margin is expected
(p.
to contribute to the available level of internal funds and hence the level of inter

1
^airfield and Yohn (2001) hypothesize that the specific mix of asset turnover and profitmargin
is not useful in predicting future profitability because the levels of these ratios are in part the product
of a firm's operating strategy. Consistent with their hypothesis, they find that it is the change in asset
turnover and change in profit margins (and not the level of asset turnover and profitmargin) that is
useful in forecasting changes in return on assets one year ahead.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
364 Journal of Financial and Quantitative Analysis

financed firm growth. an increase in profit margin reduces the

nally Conversely,
demand for external funds and contributes to a lower level of externally financed
firmgrowth.We anticipate a negative coefficient on the change in profitmargin
(ANI?NS).
Fairfieldand Yohn (2001) also note that asset turnover measures a firm's

ability to generate revenues from its assets. To the extent that a change in as
set turnover influences a firm's to generate sales from its assets in place,
ability
a change in asset turnover should contribute to realized total growth as well as

the level of internally financed firmgrowth. Recall thatexternally financed firm

growth is estimated as the difference between the realized total firmgrowth and
internally financed firm growth. Because a in asset turnover influences
change
both the realized total growth and internally financed firmgrowth, the impact of
asset turnover on financed firm growth cannot be de
externally unambiguously
termined. Hence, we do not a for the coefficient on in asset
predict sign change
turnover (Z\NS/TA).
Larger firms are more likely to grow at rates that could be financed without
access to long-term credit or to stock markets. Hence, the larger the firm size,
the smaller the firm's externally financed growth. Hence, the coefficient on the

variables, LOG_TA, is expected to be negative. In contrast, firms with greater

reliance on external as the amount of equity and debt raised
financing proxied by
as a proportion of their total assets (FIN/TA) are likely to exhibitmore externally
financed growth. Hence, the coefficient on FIN/TA is expected to be positive.
Firms with more tendto utilize external to
growth opportunities financing
pursue potentially profitable projects. To the extent that this is true, firms with
more should exhibit more financed Fol
growth opportunities externally growth.
lowing prior research (e.g., La Porta et al. (2000)), we use Tobin's Q as a proxy
for growth opportunities and measure Tobin's Q as the sum of market value of

equity plus assets minus the book value of equity deflated by total assets. The
greater the value of Tobin's Q, the greater the externally financed growth. Hence,
the expected sign for the variable Tobin's Q is positive.

B. Measurement of Externally Financed Growth

To empirically measure financed we start with Demirguc

externally growth,
Kunt and Maksimovic's (1998) implementation of a firm-based financial plan
ning model to estimate the maximum rate of growth that can be financed inter
we firm that can be achieved
nally. Specifically, compute growth by relying either
on internal cash flows or short-term For each firm, we then com
borrowing.
pute the difference between the realized rate of growth and the two measures of
constrained growth. This difference reflects the level of growth realized through
external financing.Demirguc-Kunt andMaksimovic (1998) interpretthe growth
rate the constrained rate as evidence of external of
exceeding growth financing
marginal investment.
A firm's external need depends on both the availability of internal
financing
funds as well as investment Therefore, a firm's need for external
opportunities.12

12It should also be noted that firms' internal cash flows and investment opportunities are endoge
nously determined. For example, a firm's internal cash flow depends on the profitability generated

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 365

financing must take into account both of these factors. Following Demirguc-Kunt
andMaksimovic (1998), (2002), we estimate a firm's externally financed growth
using its "percentage of sales" approach to financial Three are
planning. points
worth noting about this approach. First, the ratio of assets used in production
to sales is assumed to be constant. The here is that the total investment
upshot
required will be positively related to a firm's growth in sales. Second, the firm's

profit rate per unit of sales is assumed to be constant. we assume eco

Finally,
nomic depreciation is equal to the depreciation amount reported in firms' finan
cial statements. As a consequence, the external needs of a firm at time
financing
t can be expressed as

= *
(3)EFN, [g, A,]-[(1+g,) *(?,*/>,)],

where EFN, is a measure of external financing need, gt is firm growth at time

period t,At is firmassets at timeperiod t,bt is theproportion of thefirm's earnings

that are retained for reinvestment at time t, and Et is earnings after interest and
taxes at time t.
The on the right-hand side represents the difference between the
expression
required investment for a firm growing at gt percent less the internally available

capital for investment.

Using the model in (3), we compute two measures of constrained growth

denoted as the internallyfinanced growth rate (IG) and the short-termfinanced
growth rate (SFG). SFG represents a less conservative estimate of a firm's con
strained growth rate. The estimated IG variable represents the maximum growth
rate that can be attained if a firm strictly relies only
on its internal resources and
the payout ratio is assumed to be constant. To estimate IG, we set EFN, to zero
and compute the variable gt using equation (3). The resulting growth rate reduces
to the following equation,

= * -
(4) IG, (ROA, bt)/(l ROA, * &,),

where ROA, is the ratio of earnings after interest and taxes to assets.13 As Demir

guc-Kunt and Maksimovic (1998) point out, IG, is convex and increasing in the
firm's return on assets. This implies that greater profitability from assets in place

supports higher growth rates.

We also estimate a second constrained measure, SFG,. This esti
growth
mated growth rate represents the maximum growth rate of a firm attained through
both internal cash flows and short-term debt. The amount of short-term borrowing
undertaken is restricted such that the short-term debt to assets ratio is maintained.
The purpose of this restriction is to ensure that the growth estimate is feasible
for the underlying firms involved. The shortcoming of this assumption is that it
does not completely capture a firm's short-term borrowing capacity. The growth
estimate, SFG,, is obtained by first setting the variable bt in equation (3) to one.
from its assets in place. With respect to investments, Demirguc-Kunt and Maksimovic (1998) posit
that firms in capital intensive industries may require larger investments to grow.
13
We use earnings after interest and taxes to assets to measure how much of the net income is
generated from total assets and captures the extent of resources that are available internally within the
firm.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
366 Journal of Financial and Quantitative Analysis

This implies that the payout ratio is zero. Given thatexternal financing is limited
such that the short-term debt to asset ratio is constant and that the dividend payout
ratio is zero, the implied growth rate of thefirm is expressed in equation (5) as

= -
(5) SFG, ROLTQ/0 ROLTQ),

where ROLTQ is the ratio of earnings after interest and taxes to long-term
capital.14
Following Demirguc-Kunt and Maksimovic (1998), for each firm we calcu
late the difference between its realized sales growth rate in the year of the disclo
sure measurement andits predicted financed rate, and denote it
internally growth
as EXCESS_IG. for each firm we calculate the difference between its
Similarly,
realized sales growth rate and its predicted short-term financed rate, and
growth
denote itas EXCESS_SFG.
Following Demirguc-Kunt and Maksimovic (1998), we also compute two
metrics analogous to the two continuous metrics, EXCESS _IG and EXCESS _SFG,
to reduce the effect of outliers. for each firm we calculate the pro
Specifically,
portion of years in which its realized sales growth rate in three consecutive years
exceeds itspredicted internally financedgrowth rate, and denote itas PROP _IG.15
Similarly, for each firmwe calculate the proportion of years inwhich its realized
sales growth rate in three consecutive years exceeds its predicted short-term fi
nanced growth rate, and denote it as PROP.SFG.

C. Test Variable

The test variable, DISCL, represents the analyst evaluations of corporate

public disclosure. These scores are obtained from the annual volumes of the Re

port of theCorporate InformationCommittee (CIC) published by the Financial

Analysts Federation branch of the AIMR. To evaluate the quality of corporate

disclosures, CIC first forms subcommittees for a select group of industries. Each
subcommittee is then completed with analysts specializing in that industry.The
documents examined by each analyst within a subcommittee include annual re

ports, 10-Ks, quarterly reports, proxy statements, and other published information
such as press releases and fact books, as well as less formal disclosures
through
meetings and responses to analyst (for each selected firm within that
inquiries
industry). For uniformity of evaluation across industries as well as firms,CIC
provides a list of criteria that is to be used to evaluate disclosures. The
corporate
final disclosure scores for each a consensus of
corporation represent judgment
analysts in the industry to which the corporation Prior studies use the
belongs.
AIMR disclosure scores as a comprehensive measure of a firm's disclosure policy
(e.g., Lang and Lundholm (1993), (1996), Sengupta (1998)).

14Following Demirguc-Kunt and Maksimovic, we denote the assets of the firm not financed by
short-term debt as "long-term capital," which is obtained by multiplying a firm's total assets by one
minus the ratio of short-term liabilities to total assets.
15To examine the robustness of our results to our choice of computing proportions over a three-year
period, we repeat our analysis by computing proportions over two and four years, and find that the
three-year cut-off is not critical for our results on the relation between externally financed growth and
level of disclosure.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 367

The AIMR score is a weighted combination of scores based on evaluations

of the three aspects of a firm's disclosure policy: annual published information,

quarterly and other published information, and investor relations and related as

pects. The benefit of using the scores assigned by the analysts is that theyprovide
a ready off-the-shelf measure that is widely used in prior research as a compre
hensive measure of corporate disclosure These scores reflect
practices.16 analysts'
evaluations of the timeliness, detail, and clarity of information presented (Sen
gupta (1998)). Lang and Lundholm (1996) note that the scores quantify qualita
tive disclosure (e.g., new announcements, management discussion, and
product
analysis) and disclosure thatmay not have been reflected in published financial
statements conference calls to analysts). Therefore, our disclosure metric
(e.g.,
a evaluation of a firm's disclosure
provides comprehensive policy.
While each committee uses a standardized template to rate disclosure, they
often tailor the disclosure scores to the unique characteristics of their industries

(Botosan and Plumlee (2002)). To make AIMR scores comparable across indus
tries, we follow prior research (e.g., Healy et al. (1999), Botosan and Plumlee

(2002)) and convert raw AIMR disclosure scores towithin industry/yearranks,

defined as the rank of a given firm's total disclosure score divided by the num
ber of observations that have nonmissing values of the ranking variable. We rank
firms in ascending order, such that firms providing higher levels of disclosure
receive ranks.17 Thus, a DISCL value reflects a more
higher higher expanded
disclosure policy. As discussed previously, the higher the disclosure ranks, the
the externally financed Hence, our a positive
greater growth. hypothesis predicts
sign for disclosure rank.

D. Data

Our sample selection begins with the firm-year observations included in the
AIMR Reports dated from 1982 through 1994. We require data on Compustat
to compute variables necessary to estimate models (1) and (2) as de
regression
scribed above. Eliminating these observations from the sample selection process

yields a final sample of 1,436 firm-year observations, with a minimum of 94 ob

servations in 1987 and a maximum of 162 observations in 1990.

IV. Empirical Results

A. Descriptive Statistics

Panel A of Table 1 reports descriptive statistics for the dependent, test, and
control variables. The mean and median values of EXCESS _IG are 0.01 and

scores as a measure of firms' disclosure policies

16Examples of studies that use these disclosure
include Bamber and Cheon (1998), Botosan and Plumlee (2002), Healy, Hutton, and Palepu (1999),
Lang and Lundholm (1993), (1996), Sengupta (1998), and Welker (1995).
17Nagar, Nanda, and Wysocki (2003) note that this ranking procedure produces larger differences
between percentiles for small than for large industries. For example, in an industry with only two
firms, one will have a disclosure score of 1.0 and the other 0.0, whereas if those firms were in an
industry with many firms, their values would probably be less extreme. We replicated all our tests
are robust
using raw AIMR disclosure scores instead of the ranked disclosure scores and our results
to the use of raw disclosure scores.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
368 Journal of Financial and Quantitative Analysis

?0.01, respectively, suggesting that some sample firms grow much faster than
their internal growth rate constraint. The mean value of EXCESS JSFG is ?0.02.
The interquartile range of EXCESS _SFG suggests that there is variation in the
extent to which firms grow faster than their short-term financed growth rate. The
variable PROP-IG an estimate of the proportion of years in which a firm
provides
grows faster than our estimate of IG, the maximum internally financed growth
rate. Analogously, PROP.SFG an estimate of the proportion of years
provides
in which a firm grows faster than our estimate of SFG, the maximum short
term financed growth rate. A of the mean values of PROP JG and
comparison
PROP-SFG can highlight the relative importance of internalfunds and short-term
debt inproviding capital forgrowth. Thus, a mean value of PROP JG of 0.48 in
dicates that for 48% of firm-years, a firm's realized exceeds its estimated
growth
growth using internal funds. Of these, 11% could finance their realized growth
entirely using short-term debt.
Our testvariable, DISCL, represents the overall level of firmdisclosure and
is a rank of the total AIMR disclosure score based on a within rank
industry/year
ing. The mean value forDISCL is 0.57. The DISCL variable exhibits consider
able variation across the sample, as evidenced by the interquartile range.
On average, dividends for our sample firms constitute 3% of the total as
sets and in profit margin (Z\NI/TA) is ?0.01. The mean in asset
change changes
turnover (zANS/TA) is -0.02%. Median firm size in termsof assets is $2,612.23
million, indicating that our average sample firm is large. The firm size ranges from

$1,121.25 million in the lower quartile to over $6,118 million in the upper quar
tile. The mean and median values of Tobin's are 1.69 and 1.43, respectively.
Q
External debt and equity issuances, on average, constitute 8% of total assets.

B. Correlations

Panel B of Table 1 presents correlations vari

Spearman among independent
ables used inmodel (2). The correlations between DISCL and control variables
are low, from 0.09 to ?0.03. Consistent with the results of
relatively ranging
Lang and Lundholm (1996), disclosure policy is significantlypositively related
to firm size (LOG_TA). Although many of the pairwise correlations among the
control variables are all of the correlations are below the
significant, considerably
0.80 threshold suggested by Judge,Griffith,Hill, and Lee ((1980), p. 459) as in
dicative of a serious problem. In addition, other diagnostic measures
collinearity
indicate that collinearity is not a significant in interpreting the regression
problem
results. For example, the highest variance inflation factor (VIF) of ordinary least
squares (OLS) yearly regressions for any control variable is 2.08 (for theDIV/TA
variable) and thehighest VIF for the testvariable DISCL inany of the regressions
is only 1.06,which ismuch lower than the threshold level of 10 thatmight indicate
a (Neter, Kutner, and Wasserman
collinearity problem Nachtsheim, (1996)).,8

18The variance inflation factor (VIF) measures the interrelationship between the explanatory vari
ables and is considered preferable to examining pairwise correlations in testing for
collinearity (Neter
et al. (1996)). For each of our regressions, we evaluate the
sensitivity of our findings to influential
observations by calculating the studentized residual for each observation (Belsley, Kuh, and Welsch
(1980)). When observations with studentized residuals in excess of the absolute value of three are
dropped and the regressions reestimated, the results are similar to those reported in the paper. For

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 369

TABLE 1
Summary Statistics

All variables except EXCESSJG, EXCESS.SFG, PR0PJG, and PROP.SFG are averaged over threeyears covering the
same timespan overwhich PROP.IG and PROP.SFG are computed.
EXCESSJG = Differencebetween a firm's actual sales growthrateand itspredicted internally financedgrowthrate.
For each firm, thepredicted internally financedgrowthrate isdefinedas ROA * b/C\ ? ROA * b),
where ROA is the ratioof earnings aftertaxes and interesttoassets, and b is theproportionof the
firm's earnings thatare retainedforreinvestment.
EXCESS.SFG = Differencebetween a firm's actual sales growthrateand itspredictedshort-term financedgrowthrate.
For each firm,thepredicted short-termfinancedgrowthrate isdefined as ROLTC/(1 ? ROLTC),
where ROLTC is the ratioof earningsaftertaxand interest to long-term
capital.
PROP.IG = Proportionof years inwhich a firm's actual sales growthrate inthreeconsecutive years exceeds its
predicted internally financedgrowthrate.For each firm, thepredicted internallyfinancedgrowthrate
isdefinedas ROA * ?>/(1? ROA * b), where ROA is the ratioof earningsaftertaxes and interest to
assets, and b is theproportionof thefirm's earnings thatare retainedforreinvestment.
PROP.SFG = Proportionof years inwhich a firm's actual sales growthrate inthreeconsecutive years exceeds its
predicted short-term financedgrowthrate. For each firm,thepredicted short-term financedgrowth
rate isdefinedas ROLTC/(1 ? ROLTC), where ROLTC is the ratioof earnings aftertaxand interest
to long-term capital.
DISCL = Rank of the totaldisclosure score (obtained fromtheannual volumes of the reportof theFinancial
AnalystsFederationCorporate Information Committee)based on a within industry/year ranking.
DIV/TA = Totaldividends divided by totalassets.
NI/NS = Earnings after interest
and taxes divided by net sales.
NS/TA = Net sales divided by totalassets.
LOG.TA = Natural logof totalassets.
Tobin'sQ = ?
(Marketvalue of equity book value of equity+ totalassets)/total assets.
FIN/TA = (Issuance of equity+ issuance of debt)/totalassets.
A = Change inthevariable relativeto thepreviousyear.
Panel A. DescriptiveStatistics
Lower Upper
Variable N Mean STD Quartile Median Quartile
EXCESSJG 1,436 0.01 0.15 -0.06 -0.003 0.064
EXCESS.SFG 1,436 -0.02 0.17 -0.09 -0.027 0.046
PROP.IG 1,436 0.48 0.34 0.33 0.33 0.67
PROP.SFG 1,436 0.37 0.34 0.00 0.33 0.67
DISCL 1,436 0.57 0.24 0.37 0.58 0.77
DIV/TA 1,436 0.03 0.02 0.01 0.02 0.03
AEBHfTA 1,436 -0.01 0.04 -0.01 0.00 0.01
A NS/TA 1,436 -0.02 0.07 -0.05 -0.01 0.03
TA ($ inmill.) 1,436 5,083.77 6,768.36 1,121.25 2,612.23 6,118.10
Tobin'sQ 1,436 1.69 0.78 1.16 1.43 1.98
FIN/TA 1,436 0.08 0.08 0.03 0.05 0.10
Panel B. Spearman Correlation
Variable DISCL DIV/TA LOG.TA Tobin'sQ FIN/TA

DIV/TA 0.04
?1EBIT/TA -0.01 0.08***
A NS/TA -0.03 -0.03 0.11***
LOG.TA 0.09*** 0.03 0.01 -0.09***
Tobin'sQ 0.04* -0.23*** 0.18*** -0.07** -0.05*
FIN/TA 0.07*** -0.16*** -0.11*** -0.16*** 0.04* -0.14***
PROP.IG 0.11*** -0.52*** -0.04 0.16*** -0.04 -0.22*** 0.20***
PROP.SFG 0.05** -0.56*** -0.09*** 0.17*** -0.01 -0.35*** 0.31***
***
,**, significantat the 10%, 5%, or 1% level(two-tailedtest), respectively.

C. Regression Results

Table 2 presents 2SLS regression results using PROP JG as the dependent

variable.19 We do not use a pooled, time-series, cross-sectional model because in
a the /-statistics may be biased due to positive cross
pooled regression upward

each of our regressions, the null hypothesis of homoskedasticity could not be rejected using theWhite
(1980) test at the 0.10 level. To check whether the relation between our test variable DISCL and the
dependent variable is nonlinear, we plotted the residuals against DISCL but the plots did not show a
nonlinear pattern.
19Instrumental variables estimation of DISCL to extract the endogenous effect of disclosure yields
an average adjusted R2 of 0.43 for the 11 years.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
370 Journal of Financial and Quantitative Analysis

correlations in the residuals.20 Rather, we estimate annual cross-sectional regres

sions from 1984 through 1994 and use the estimated slope coefficients for the
11 yearly regressions to obtain an across-year mean coefficient for each variable.
Consistent with the prior literature (Bernard (1987)), tests of significance for the
time-series mean parameter estimates are based on Zl and Z2 statistics.21
The adjusted R2s formodel (2) range from 0.15 to 0.39. All yearly co
efficients of DIV/TA are negative, indicating thathigh dividends are associated
with lower rates of externally financed growth. all the yearly coeffi
Similarly,
cients of Z\NS/TA are positive, that rates of growth that exceed the
suggesting
predicted internallyfinanced growth rate correlate positively with the change in
asset turnover. All but four of the yearly coefficients for ZANI/NS are negative,
suggesting thatfirmswith larger increases in profitmargin are less likely to grow
at rates that require them to obtain external Six yearly coefficients for
financing.
LOG-TA are The on LOG_TA indicates that firms
negative. negative sign large
tend to grow at rates that can be financed without access to external In
financing.
contrast, all but two of the yearly coefficients of Tobin's Q and all but one of the

yearly coefficients of FIN/TA are positive and in the expected direction. For most
years, the coefficients on Tobin's Q and FIN/TA are also statistically significant,
indicating that the growth opportunities and reliance on external are
capital posi
tivelyassociated with rates of growth thatexceed thepredicted internallyfinanced
growth rate. Both theZl and theZ2 test statistics indicate thatDIV/TA, Z\NS/TA,
LOG-TA, Tobin's Q, and FIN/TA are statistically significantat the 0.05 level.
Our main interest, in terms of HI, is whether DISCL is positively associated
with the rates of growth that exceed the rate predicted by the use of internal funds.
All yearly coefficients of DISCL are positive. For six years, these coefficients are
also statistically significant.Both theZl and theZ2 test statistics indicate that the
coefficienton DISCL is statistically significantat the0.01 level.22 The implication
is that an expanded firm disclosure policy is positively associated with rates of
growth that require external financing. In other words, our is consistent
finding
with the notion that disclosure affects firm growth access to lower
by providing
cost external financing.
Table 3 presents 2SLS a
results on both
regression yearly basis and an across
years significance test, using EXCESS_SFG
the asvariable. Recall
dependent
that EXCESS-SFG is the excess of a firm's realized sales rate and its
growth
predicted short-term financed growth rate. The across-years tests
significance

20The untabulated results for the pooled regression are similar to the results for the cross-sectional
regressions reported in this paper.
21Z1 and Zl statistics testwhether the time-series mean f-statistic equals zero,

Zl = -? I V-^ -? t;
Jt =
and Z2
V^V
> - -,
ytt y/kj/ikj 2) stddeviO/v^V^T)
where t is the i-statistic for year j, k is the
degrees of freedom for year j, and TV is the number of
years. Zl assumes independence in the annual ?-statistics whereas Z2 corrects for potential lack of
independence.
22We also test for an alternative specification in which FIN/TA is interacted with DISCL. The
coefficient on the interaction term is positive and statistically significant at the 0.01 level,
suggesting
that firms that disclose more and access external markets exhibit higher
externally financed growth
rates.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 371

TABLE 2
Regression Results
Dependent Variable: Firms Growing Faster than Predicted Internally Financed Growth Rate

= a + fr DISCL,, + ?2U\\Jit/lAit + ?3AM\it/MSit + ?4AMSit/MFAit

PROP.IG/
+ /35LOG_TA/f+ /36Tobin's Qit + ?7F\Nit/TAj + eit

Regressions are estimated using two-stage leastsquares. The instrumentalvariables forDISCL are lagged DISCL, log
ofmarketvalue of equity,historicalcorrelationbetween annual returnsand earnings computed over thepreceding five
years, returnon assets, and standard deviationof annualmarket-adjustedstock returns.All independentvariables are
averaged over threeyears covering thesame timespan overwhich PROP.IG iscomputed. Mean parameterestimates
estimates fromtheyearlyOLS regressionstoobtain an across-yearmean and f-statistics.
use the 11 coefficient Z1 and
mean i-statistic
Z2 statisticstestwhether the time-series equals zero,

V? 2)
\lk,/^i
where t is the i-statisticforyear y,k is thedegrees of freedomforyear j, and N is thenumberof years.
PROP.IG = Proportionof years inwhich a firm's actual sales growthrate inthreeconsecutive years exceeds itspre
dicted internally financedgrowthrate. For each firm,the predicted internally financedgrowth rate is
definedas ROA * fc>/(1 ? ROA *
b), where ROA is the ratioofearningsaftertaxes and interest
toassets,
and b is theproportionof thefirm's earnings thatare retainedforreinvestment.
DISCL = Rankof the totaldisclosure score (obtained fromtheannual volumes of the reportof theFinancialAnalysts
FederationCorporate Information Committee)based on a within industry/year ranking.
DIV/TA = Totaldividendsdivided by totalassets.
NI/NS = Earnings after interest
and taxes divided by net sales.
NS/TA = Net sales divided by totalassets.
LOG.TA = Natural logof totalassets.
Tobin'sQ = ?
(Marketvalue of equity book value of equity+ totalassets)/total assets.
FIN/TA = (Issuance of equity+ issuanceof debt)/totalassets.
A = Change inthevariable relativeto thepreviousyear.
ParameterEstimates
(i-statisticinparentheses)

Year Adj. R? Intercept DISCL DIV/TA Z\EBIT/NS Z^NS/TA LOG.TA Tobin'sQ

1984 1.179 0.0963 -5.3476 0.1316 0.4585 -0.0769 -0.0838 0.0278

(4.32)*** (0.76) (-3.57)*** (0.30) (1.15) (-2.35)*** (-1.44)* (0.08)
1985 0.8517 0.3501 -9.3138 -0.6349 0.4358 -0.0747 0.1301 0.0303
(3.43)*** (2.80)*** (-5.86)*** (-0.85) (1.10) (-2.53)*** (1.90)** (0.20)
1986 0.7945 0.0954 -10.1179 0.6987 0.7701 -0.0334 0.1019 0.1528
(3.65)*** (0.92) (-7.00)*** (1.19) (2.22)** (-1.38)* (2.14)** (1.53)*
1987 0.3771 0.1696 -7.8369 -3.4381 0.5267 0.0101 0.0699 0.1813
(1.56) (1.37)* (-4.33)*** (-2.35)*** (1.58)* (0.37) (1.42)* (0.66)
1988 0.7941 0.1293 -11.5766 -1.4829 0.7819 -0.0359 0.1457 0.6031
(3.91)*** (1.20) (-6.05)*** (-1.69)* (2.59)*** (-1.58)* (2.67)** (2.28)***
1989 0.5917 0.2226 -6.5107 -1.8273 0.8505 0.0018 -0.0318 0.6591
(3.01)*** (2.06)** (-5.17)*** (-1.57)* (2.82)*** (0.08) (-0.72) (2.87)***
1990 0.5801 0.0761 -9.9954 -1.0498 0.6699 0.0114 0.0398 0.5044
(3.10)*** (0.75) (6.16)*** (-1.17) (2.33)*** (0.53) (1.02) (1.48)*
1991 0.3247 0.0621 -9.9909 1.2545 0.7784 0.0228 0.0759 0.8866
(1.72)** (0.59) (-6.20)*** (0.60) (1.45)* (1.11) (1.32)* (1.97)**
1992 0.3674 0.1315 -8.9696 0.5229 0.4443 0.0222 0.0423 0.4111
(2.19)** (1.53)* (-5.82)*** (1.48) (1.96)** (1.01) (2.07)** (2.51)***
1993 0.6009 0.1504 -10.1417 -1.3336 0.2142 -0.0126 0.0529 0.3886
(3.32)*** (1.64)** (-7.09)*** (-1.56)* (0.49) (-0.59) (1.66)** (1.32)*
1994 1.0973 0.1529 -10.1638 -0.5937 1.1054 -0.0468 0.0111 -0.1556
(5.09)*** (1.52)* (-6.18)*** (-1.21) (2.21)** (-1.91)** (0.28) (-0.83)
Mean 0.6871 0.1488 -9.0877 -0.7048 0.6396 -0.0193 0.0504 0.3354
Z1 (9.71)*** (4.53)*** -17.45)*** (-1.71)** (5.48)*** (-1.99)** (3.39)*** (3.87)***
Z2 (10.35)*** (6.69)*** -18.94)*** (-1.49)* (8.76)*** (-1.69)** (3.06)*** (3.88)***
***
at the 10%, 5%, or 1% level(one-tailedtestexcept where thesign isnotpredicted), respectively.
significant

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
372 Journal of Financial and Quantitative Analysis

inTable 3 indicate that all control variables except LOG_TA and Tobin's Q are
associated with rates of growth thatexceed short-termdebt financed growth.Our
main interest is inwhether a higher level of disclosure is associated with realized
rates of growth that exceed the estimated rate of growth internal funds and
using
short-term debt. Hence, our focus is on the DISCL variable. If, in fact, disclosure
policy allows firms to better pursue potentially profitable projects using external
funds, then we should observe a relation between the level of disclosure
positive
and externally financed growth. The test variable DISCL is significantwith the
predicted positive sign indicating thata higher disclosure level is associated with
more financed
externally growth.
Lang and Lundholm (2000) suggest that, for some firms, disclosure can have
a more insidious in that it helps
effect less deserving firms to window-dress and

hype their stock price and thereby issue new capital at a lower cost. Therefore, a

potential concern with the observed relation between financed

positive externally
growth and disclosure is that it does not distinguish between firms that access
external fundsto grow faster because their disclosure them access to capital
gains
markets, or because disclosures to hype the stock to attain lower
they manage
cost equity capital. However, the "hyping" should bias against
argument finding
a relation between an
externally financed growth rate and disclosure.
To shed light on this issue, we partition our
sample into two groups such that
the incentives to manage disclosures vary across the two groups. Theory suggests
thatexternal financing is particularly important ifa firmhas growth opportunities
outstanding and if these growth opportunities cannot be financed sufficientlyby
a firm's internal funds. Toward this end, we first identify those
sample firms
that issued equity in a given year.23 Following Korajczyk and Levy (2003), we
then classify these sample firms into financially constrained and unconstrained

subsamples using themedian value of Tobin's Q and reestimate model (2) for
each subsample. Our expectation is that the positive relation between externally
financed growth and disclosure should be more for our
pronounced financially
constrained
subsample.24
When we use PROP.IG as thedependent variable, the coefficient on DISCL
for thefinancially constrained subsample is higher than the coefficient on DISCL
for thefinancially unconstrained subsample (0.3065 versus 0.2111 ) and thediffer
ence is statistically significantat the0.01 level.When we repeat the
analysis with
PROP-SFG as thedependent variable, the coefficienton DISCL for thefinancially
constrained subsample is higher than the coefficient on DISCL for thefinancially
unconstrained subsample (0.1228 versus 0.0406), and the difference is sta
again

23Our focus on this set of firms is based on prior research that finds that firms manage earn
ings/disclosure prior to raising funds through external equity issuance. For example, Teoh, Welch,
andWong (1998a), (1998b) show that firmsmanage their
earnings upward to boost share prices prior
to initial public offerings (IPO) as well as
secondary equity offerings (SEO), and that the stock price
corrects itself in the periods following the IPO and SEO. In contrast, Lang and Lundholm (2000)
focus on disclosure and find that firms thatmaintain a consistent disclosure level do not exhibit an
unusual return behavior relative to the control firms subsequent to an equity
offering announcement.
However, firms that altered their disclosure policy to "hype" their stock prior to equity issuance suffer
a decline in value in the period
following the equity issuance.
24Financially constrained firmswith a higher level of disclosure should be able to realize more ex
ternally financed growth. Furthermore, for financially constrained firms, the incentive to use disclosure
as a means to obtain lower cost external
financing should dominate any other competing incentives.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 373

TABLE 3
Regression Results
Variable: Firms Growing Faster than Predicted Short-Term Financed
Dependent
Growth Rate

= oi + ?^ DISCL/? + ?2DW/jt/TAit + ?3AM\it/MSit + ?4AMSit/NFAit

PROP.SFG,
+ /35LOG_TA,y + ?6Tobin's Q/f + ?7F\Hlt/JAit + eit

Regressions are estimated using two-stage leastsquares. The instrumentalvariables forDISCL are lagged DISCL, log
ofmarketvalue of equity,historicalcorrelationbetween annual returnsand earnings computed over thepreceding five
years, returnon assets, and standard deviationof annual market-adjustedstock returns.All independentvariables are
averaged over threeyears covering thesame timespan overwhich PROP.SFG iscomputed. Mean parameterestimates
use the 11coefficientestimates fromtheyearlyOLS regressionstoobtainan across-yearmean. Z1 and Z2 statisticstest
whether the time-seriesmean i-statistic
equals zero,

f
_i_
?-Y - stddev(t)/VN
v^tt 2)
y/kj/(kj
where t isthe i-statisticforyear j, k is thedegrees of freedomforyear /,and N is thenumberof years.
PROP.SFG Proportionof years inwhich a firm'sactual sales growthrate inthreeconsecutive years exceeds its
predicted short-termfinanced growthrate. For each firm,the predicted short-termfinancedgrowth
rate isdefinedas ROLTC/(1 ? ROLTC), where ROLTC is the ratioof earnings aftertaxand interestto
long-term capital.
DISCL Rank of the totaldisclosure score (obtained fromthe annual volumes of the reportof the Financial
AnalystsFederationCorporate Information Committee)based on a within industry/year
ranking.
DIV/TA Totaldividends divided by totalassets.
NI/NS Earnings after interest
and taxes divided by net sales.
NS/TA Net sales divided by totalassets.
LOG.TA Natural logof totalassets.
?
Tobin'sQ (Marketvalue of equity book value ofequity+ totalassets)/total assets.
FIN/TA (Issuance of equity+ issuance of debt)/totalassets.
A Change inthevariable relativeto thepreviousyear.
ParameterEstimates
(i-statisticinparentheses)

Year Adj. R2 N Intercept DIV/TA Z^EBIT/NS zANS/TA LOG.TA Tobin'sQ FIN/TA

1984 0.6949 -0.0617 -3.6256 -0.2291 0.6204 -0.0271 -0.0855 0.5142

(3.01)*** (-0.57) (-2.86)*** ( -0.62) (1.88)** (-0.98) (-1.74)** (1.16)
1985 0.7487 0.0886 -6.4748 -0.8587 0.7193 -0.0555 0.0408 0.4731
(3.40)*** (0.79) (-4.61)*** ( -1.30)* (2.05)** (-2.11)** (0.68) (1.55)*
1986 0.5888 0.0306 -7.5416 0.1396 0.8654 -0.0114 -0.0016 0.8086
(2.96)*** (0.32) (-5.64)*** (0.32) (2.73)*** (-0.52) (-0.04) (3.77)***
1987 0.1889 0.1785 -5.1008 -3.2618 1.1638 0.0282 -0.0385 0.9449
(0.89) (1.69)** (-3.26)*** ( -2.56)*** (4.03)*** (1.21) (-0.90) (5.07)***
1988 0.8047 0.0901 -8.6698 -0.8331 0.8177 -0.0354 0.0029 0.7571
(3.75)*** (0.80) (-4.49)*** ( -0.73) (2.58)*** (-1.48)* (0.05) (3.23)***
1989 0.7216 0.1236 -4.3919 -1.4023 0.6997 -0.0243 -0.0891 1.0498
(3.35)*** (1.05) (-3.44)*** ( -1.10) (2.12)** (-0.99) (-1.86)** (2.71)***
1990 0.5184 0.0621 -8.5861 -1.1249 0.7737 0.0031 -0.0078 0.8602
(2.59)*** (0.59) (-5.12)*** ( -1.20) (2.57)*** (0.14) (-0.19) (2.24)**
1991 0.1564 0.0194 -9.5676 1.5756 0.9883 0.0394 0.0606 0.6382
(0.81) (0.19) (-5.64)*** (1.87)** (2.50)** (1.74)** (1.66)** (1.89)**
1992 0.0684 0.0433 -39.0379 -0.3854 0.6904 0.0519 0.0302 0.9022
(0.40) (0.51) (-6.21)*** (-0.44) (2.25)*** (2.57)*** (0.94) (2.98)***
1993 0.3622 0.0942 -9.1072 -1.5232 0.3403 0.0047 0.0271 1.4795
(1.96)** (1.00) (-5.98)*** ( 1.73)** (0.75) (0.21) (0.82) (3.69)***
1994 0.9064 0.0966 -9.5296 -0.4495 0.9209 -0.0304 -0.0139 0.3916
(3.77)*** (0.87) (-5.20) (-0.82) (1.64)** (-1.12) (-0.32) (1.01)
Mean 0.5236 0.0696 -10.1484 -0.7593 0.7818 -0.0052 -0.0068 0.8018
Z1 (7.39)*** (2.17)** (-14.43)*** ( 2.28)** (6.90)*** (-0.36) (-0.24) (8.05)***
Z2 (6.84)*** (3.63)*** (-14.39)*** ( 2.29)** (9.82)*** (-0.28) (-0.26) (7.41)***
,***significant
at the 10%, 5%, or 1% level(one-tailedtestexcept where thesign isnotpredicted), respectively.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
374 Journal of Financial and Quantitative Analysis

tistically significant at the 0.10 level. Taken these results suggest that
together,
while increased disclosure may access to lower cost external fi
activity improve
nancing, the impact of disclosure ismore pronounced for firms forwhich lower
cost external financing is important.

D. Sensitivity Tests

We subject our regression results to a number of robustness checks in order to

address potential concerns about estimation and other model is
misspecification
sues. An our tests is that the firm's profit rate on
assumption underlying empirical
marginal sales equals its average profit rate. This also implies that the firm's cost
structure remains constant. Following Demirguc-Kunt and Maksimovic (1998),
we the growth rate estimates to allow for a lower rate of earnings on new
modify
growth. we introduce a parameter z that measures the ratio of the
Specifically,
rate on new sales to the firm's average a modified
profit profit rate to derive IG
and a modified SFG rate given by

= - z*
(6) IG, (ROA, * bt)/{\ ROA, * bt) and
=
(7) SFG, ROLTC,/(l-z*ROLTC,).

As a specification we
reestimate
check, the specifications in Table 4 for z = 0,

0.25, 0.50, 0.75, and 1. Table

4 presents the regression results assuming different
values of z.25 To save space, we only report the results for the estimated DISCL
coefficients that are returned after we impose changes to the computation of our
variable for given values of z. Again, we report the cross-sectional
dependent
mean using theestimated coefficients forDISCL for the 11 yearly regressions.We
test significance for the mean parameter estimate based on Zl and Z2 statistics.
An inspection of the results inTable 4 indicates that theassumption of theequality
of the profit rate on marginal sales and the average profit rate is not crucial for our
results on the relation between disclosure and externally financed growth rates.
Another assumption our measurement of financed
underlying externally
growth constraints is that a firm's asset turnover remains constant. To the ex
tent that this assumption does not hold, the measure of externally financed growth
used in the studywill reflectboth internallyfinanced growth (through improved
operating performance) and externally financed growth. Therefore, we modify
the estimates of the growth rate to allow for a higher rate of asset turnover on new

growth. we introduce a parameter y that measures the ratio of the

Specifically,
assets turnover on the new sales to the firm's average assets turnover to derive a
modified IG and a modified SFG rate given by

= -
(8) IG, (ROA, * bt)/{\ * y ROA, * bt) and
=
(9) SFG, ROLTC,/(l*y-ROLTC,).
25To assess the impact of lower cost structure (and higher profitability), we also reestimate the
specification inTable 4 for z= 1.05 and z= 1.10. Note that in the presence of market competition, there
is less room for a firm to adjust its cost structure to improve its profitmargin. Therefore, we specify
a narrower range of the parameter values of z. Unreported results for these alternative specifications
show no evidence of a differing association between disclosure and externally financed growth for our
sample.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 375

TABLE 4
Sensitivity Analyses forRegressions of Short-Term Financed Growth Rate Based on
Different Marginal Profit Rate Assumptions

PROPJG/(orPROP-SFG) = a + frDISCL/f + ?2D\y it/JAit+ ?3AN\it/NSit

+ ?4ANSjt/NFAjt + /35LOG_TA/f
+ /36Tobin's Qit + ?7F\b\it/JAit + eit

Only theparameterestimates returnedforDISCL are reported.Regressions are estimatedusing two-stage leastsquares.

The instrumentalvariables forDISCL are laggedDISCL, logofmarketvalue ofequity,historicalcorrelationbetween annual
and earningscomputed over thepreceding fiveyears, return
returns on assets, and standarddeviationofannualmarket
All independentvariables are averaged over threeyears covering thesame timespan overwhich
adjusted stock returns.
PROP-IG and PROP.SFG are computed. Mean parameterestimates use the 11 coefficient estimates toobtainan across
yearmean. Parametervalues of z measure the ratioof theprofitrateon new sales to thefirm'saverage profitrate.Z^ and
Z2 teststatisticstestwhether the time-series
mean f-statistic
equals zero,

Z1 ? _1_ '
Y - - 1
stddev(i)/\/A/
VTVytl y/kj/(kj 2)
foryear j, k is thedegrees of freedomforyear _/',
where t isthe f-statistic and N is thenumberof years.
PROP.IG = Differencebetween a firm's actual sales growthrateand itspredicted internally financedgrowthrate.
For each firm,thepredicted internally financedgrowthrate isdefinedas ROA * ?>/(1 ? z * ROA * b),
where ROA is the ratioof earningsaftertaxes and interest toassets, and b is theproportionof thefirm's
earnings thatare retainedforreinvestment.
PROP-SFG = Proportionof years inwhich a firm's actual sales growthrate inthreeconsecutive years exceeds its
predicted short-term financedgrowthrate.Foreach firm, thepredicted short-term financedgrowthrate
isdefined as ROLTC/(1 ? z * ROLTC), where ROLTC is the ratioof earnings aftertaxand interestto
long-termcapital.
DISCL = Rank of the totaldisclosure score (obtained fromthe annual volumes of the reportof the Financial
AnalystsFederationCorporate Information Committee)based on a within industry/year ranking.
DIV/TA = Totaldividends divided by totalassets.
NI/NS = Earningsafter interest and taxes divided by net sales.
NS/TA = Net sales divided by net fixedassets.
LOG-TA = Natural logof totalassets.
Tobin'sQ = ?
(Marketvalue of equity book value of equity+ totalassets)/total assets.
FIN/TA = (Issuance of equity+ issuance of debt)/totalassets.
A = Change inthevariable relativeto thepreviousyear.
Dependent Variable
PROP.IG PROP-SFG

Mean Mean
Parameter Z1 Z2 Parameter Zl Z2

Z? 1 0.1488 (4.53)*" (6.69)*** 0.0696 (2.17)** (3.63)***

z= 0.75 0.1118 (3.96)*** (7.85)*** 0.0491 (2.00)** (4.23)***
z= 0.50 0.1117 (3.95)*** (7.60)*** 0.0571 (2.36)*** (5.01)***
z= 0.25 0.1168 (4.15)*** (9.04)*** 0.0503 (2.09)** (5.00)***
z= 0 0.1076 (3.77)*** (8.41)*** 0.0751 (2.51)*** (9.34)***
'**
at the 10%, 5%, or 1% level(one-tailedtestexcept where thesign isnotpredicted), respectively.
significant

As a specification check, we reestimate the specifications in Tables 2 and 3

fory= 1.05 and y= 1.10.26 Results (not reported) of regressionmodel (2) indicate
that the assumption of the equality of the ratio of asset turnover on new sales to the
firm's average assets turnover is not crucial for our results on the relation between
disclosure and externally financed growth rates.
As we note in Section II, to the extent that our sample firms rely more on

public debt or equity instead of private debt, the effectof disclosure on externally
financed growthmay be more pronounced forfirms issuing public debt or equity.
As a sensitivity test,we identifythose sample firms that issued either private debt

26Our choice of the parameter values of y are based on anecdotal evidence that suggests that asset
turnover tends to be sticky.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
376 Journal of Financial and Quantitative Analysis

or public debt and equity in a given year from the Securities Data Company New
Issues database and classify the sample firm-years into two groups?those issuing
public debt or equity and those issuing private debt.We then reestimate regression
model (2) for each subsample using PROPJSFG as the dependent variable. The
magnitude of the coefficient on DISCL forfirms that issued public debt or equity
in a given year is 0.172 and is statistically significant at the 0.01 level. For firms
that issued only private debt in a given year, the magnitude of the coefficient on
DISCL is 0.113 but is statistically insignificant at the 0.10 level. These results
are consistent with the notion that disclosure is of more consequence to firms that
raise funds in public capital markets.
As an alternative to the 2SLS estimation, we also address the endogeneity of
a firm's disclosure policy in our empirical tests bymeasuring DISCL prior to the
time period over which economic growth of the firm is computed. In theseOLS
estimations, we also average other firm characteristics in model (2) over the same
time period for which firm growth is computed. As an example, for a firm-year
observation with a fiscal year ending on December 31,1990, we use DISCL for
the year 1990, compute externally financed growth during 1991, 1992, and 1993,
and average other firm characteristics (e.g., DIV.TA) for the years 1991, 1992,
and 1993. results of this alternative specification are very similar to
Unreported
those reported inTables 2 and 3 where disclosure and externally financed growth
metrics are measured over contemporaneous time periods. Overall, the attempt to

explicitly address self-selection bias supports the hypothesis that analyst rankings
of overall firmdisclosure are positively associated with firmgrowth supported by
external financing.
Thus far we have focused on the level of the disclosure scores. To ensure

that our results are not driven correlations, we also examine the asso
by spurious
ciation between changes in our excess growth metrics and changes in disclosure

scores.27 We focus on
"significant and sustained
improvements" in disclosure
for several reasons.28 First, a significant and sustained in disclo
policy change
sure policy is more a deliberate
likely to reflect managerial policy. Second, while
extant theory is quite clear on the effects following disclosure improvements, it

provides littleguidance with respect to the impact of a decline indisclosure policy.

a decline ismore
Furthermore, in disclosure likely to be endogenous to changes in
disclosure costs that serve to moderate firms' disclosure policies. As Lang (1999)
points out, a time-series approach works best if the disclosure changes involved
are
"exogenously imposed."

27We thank the reviewer for suggesting this robustness test.

28A concern in using changes in disclosure scores is that analysts' evaluations may be influenced
by considerations other than improvements in firm disclosure policy. To address this concern, Healy
et al. (1999) examine the reasons offered by analysts in assigning improved scores for the subject
firms involved. Examples of these reasons include: "(a) Improved segment disclosures; (b) more in
depth discussion of operations and financial performance, and more candid management discussion of
the company's prospects in annual and quarterly reports; (c) publication of supplemental disclosures
in fact books; and (d) improved investor relations through increased analyst access to top manage
ment and additional company meetings and presentations for analysts" (Healy et al. (1999), p. 489).
This supports the contention that analysts upgrade their evaluations based on improvements in firms'
disclosure policies.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 377

et al. (1999), we compute the change in average relative

Following Healy
ranking (Z^DISCL) forour sample firms as
-2 -4
= - DISCL DISCL.
(10) ADlSCUt --^
]T
r=0 f=-3

For each firm,we identify the year during the sample period forwhich the
largest increase in average relative ranking takes place and focus on these firm

years to examine the relation between the changes in our excess growth metrics
and Z\DISCL. The reduced sample consists of 180 firmswith an average change
of 16.1 percentage No firm enters the sample more than once and the
points.
sample period ranges from 1986 to 1994. We then compute AEFGit in a manner
analogous toADlSCLit.

We, therefore,consider the following specification inwhich all independent

variables are as flow(change) variables,
incorporated

= a + ?i ADlSCLit + /^DIV^/TA-, + ?3ANlit/NSit

(11) AEFGit
+ ?4ANSit/NFAit + ?5ALOG TAf/+ /^Tobin's Q/f
+ #FIN/,/TA/,,

where all variables are as defined before and the change operator (A) for the
and independent variables other than DISCL represents a change in the
dependent
= Tobin's ?
to the previous
variable relative year.29 For example, A Tobin's Qit Q/f
Tobin's Qit-x
In Table 5, we report the results of estimating regression equation (11). The
firstmodel reports results using Z\EXCESS_IG as thedependent variable to proxy
for Z\EFG and the second model uses Z\EXCESS _SFG; as thedependent variable.
The tenor of our results is essentially unaltered by this specification. Specifically,
the coefficient on Z\DISCL is positive and statistically significantat the0.01 level
in both models, suggesting that changes in disclosure levels are associated with
changes in externally financed growth. Overall, the evidence from the change
model complements the level specification, highlighting the importance of dis
closure in easing access to external financing.

V. Conclusion
Extant research posits that information asymmetry and agency conflicts ad

versely affect the ability of firms to pursue potentially profitable projects. More
specifically, these factors are conjectured to increase the cost of external financ

ing. A related literature posits that an expanded and credible disclosure policy
will serve to improve the ability of firms to fund their growth opportunities. In
part, an expanded disclosure serves to reduce information asymmetry and
policy
mitigate agency conflicts by improving the ability of investors tomonitor inside
managers. As a consequence, it is posited that disclosure will positively affect a

firm's growth rate by improving firm access to lower cost external financing.

29We also compute the change variables over the same time period as that used for theDISCL, and
our inferences are qualitatively similar to those reported in the paper.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
378 Journal of Financial and Quantitative Analysis

TABLE 5
Regression Results for the Change Model

= a + ?iADlSCLjt + ?2D\Vit/lkit +
(11) AEFGit ?3AH\it/MSit + ?4ANSit/JAit
+ ?5ALOG TA/, + /36Z\Tobin's Qit + /?7FIN,Y/TA?

Coefficientson theyear dummiesare not reported.

EFG Externallyfinancedgrowthproxiedby EXCESS.IG or EXCESS-SFG.
EXCESSJG Differencebetween a firm's actual sales growthrateand itspredicted internally financedgrowth
rate.Foreach firm, financedgrowthrate isdefinedas ROA * b/ ( 1- ROA *
thepredicted internally
b), where ROA is the ratioof earningsaftertaxes and interesttoassets, and b is theproportion
of
thefirm's
earnings thatare retainedforreinvestment.

Z\EXCESS_IG - - - EXCESS-IG.
y^EXCESSJG
EXCESS-SFG Differencebetween a firm'sactual sales growthrateand itspredictedshort-term financedgrowth
rate. For each firm,the predicted short-termfinancedgrowth rate is defined as ROLTC/(1 ?
ROLTC), where ROLTC is the ratioof earningsaftertaxand interest
to long-termcapital.
?AEXCESS-SFG - EXCESS-SFG EXCESS-SFG.
0Y t=0 ? J2
.--?
DISCL Rank of the totaldisclosure score (obtained fromtheannual volumes of the reportof theFinancial
AnalystsFederationCorporate Information Committee)based on a withinindustry/year ranking.
DIVATA Totaldividendsdivided by totalassets.
zANI/NS Change inearningsafter interest and taxes divided by netsales relativetopreviousyear.
ENSATA Change innet sales divided by totalassets relativetopreviousyear.
A LOG-TA Change innatural logof totalassets relativetopreviousyear.
zATobin'sQ Change inTobin's Q relative to previous year where Tobin's Q is (marketvalue ofequity ?
book value of equity+ totalassets)/total assets.
FIN/TA (Issuance of equity+ issuance ofdebt)/totalassets.
-2 1 -4
/ADISCL> -1 D|SCL? DISCL
3y^r=0 2 X!
f=-3
ParameterEstimates
(f-statisticinparentheses)

Dependent Variable
Variable zAExcess.lG Z^Excess.SFG

Intercept 0.0429 0.1021

(0.68) (1.38)
ZADISCL 0.0476 0.0693
(1.37)* (1.70)**
DIVATA -2.5932 -2.9291
(-4.41)*** (-4.24)***
AEBHfTA -0.8647 -1.5705
(1.83)** (-2.83)***
ANS/TA 0.2137 0.2048
(1.60)* (1.31)*
A LOG-TA -0.0059 -0.0098
(-0.79) (-1.10)
ZXTobin's
Q 0.0292 0.0012
(2.15)** (0.70)
Financing/TA 0.2883 0.4069
(2.61)*** (3.14)***
Year dummy Included Included
N 147 147
Adj. R 0.19 0.26
***
at the 10%, 5%, or 1% level(one-tailedexcept where thesign isnotpredicted), respectively.
**, significant

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 Khurana, Pereira, and Martin 379

To evaluate
this prediction, we examine the association between
empirically
a firm's and its disclosure policy. In particular, we investigate whether
growth
an expanded disclosure allows firms to grow at rates exceeding those that
policy
could be attained on their internal sources of funds or short
by relying strictly
termborrowing.We firstestimate a predicted growth rate foreach firm if it strictly
relies on internal cash flows or short-term borrowing. We next compute the differ
ence between realized firm growth and the predicted rate. This measure
growth
reflects the amount of growth supported through external finance. Using a cross
section of U.S. firms, we find a relation between a firm's fi
positive externally
nanced growth rate and its level of disclosure. Our results are robust to alternative
measures of externally financed of disclosure, and the inclu
growth, endogeneity
sion of other firm-specific factors relevant in this setting.

References
Admati, A., and P. Pfleiderer. "Forcing Firms to Talk: Financial Disclosure Regulation and Externali
ties." Review of Financial Studies, 15 (2000), 479-519.
Almeida, H., and M. Campello. "Financial Constraints and Investment-Cash Flow Sensitivities: New
Research Directions." Working Paper, New York Univ. and Univ. of Illinois (2002).
Alti, A. "How Sensitive is Investment to Cash Flow When Financing is Frictionless?" Journal of
Finance, 58 (2003), 707-722.
Association for Investment Management and Research. Corporate Information Committee Report of
Association for Investment Management and Research Corporate Information Committee. New
York, NY: AIMR, 1981-1995.
Bamber, L., and Y Cheon. "Discretionary Management Earnings Forecast Disclosure: Antecedents
and Outcomes Associated with Forecast Venue and Forecast Specificity Choices." Journal of Ac
counting Research, 36(1998), 167-190.
Belsley, D.; E. Kuh; and R. Welsch. Regression Diagnostics, Identifying Influential Data and Sources
of Collinearity. New York, NY: JohnWiley (1980).
Bernard, V. "Cross-Sectional Dependence and Problems in Inference inMarket-Based Accounting
Research." Journal of Accounting Research, 25 (1987), 1^48.
Bhattacharya, S., and G. Chiesa. "Proprietary Information, Financial Intermediation, and Research
Incentives." Journal of Financial Intermediation, 4 (1995), 328-357.
Botosan, C. "Disclosure Level and the Cost of Equity Capital." Accounting Review, 72 (1997), 323
349.
Botosan, C, and M. Plumlee. "A Reexamination of Disclosure Level and the Expected Cost of Equity
Capital." Journal of Accounting Research, 40 (2002), 21^4-0.
Bushman, R., and A. Smith. "Financial Accounting Information and Corporate Governance." Journal
of Accounting and Economics, 32 (2001), 237-333.
Bushman, R., and A. Smith. "Transparency, Financial Accounting Information, and Corporate Gov
ernance." FRBNYEconomic Policy Review, 9 (2003), 65-87.
Campbell, S. "Optimal Investment Financing Decisions and theValue of Confidentiality." Journal of
Financial and Quantitative Analysis, 14 (1979), 913-924.
Cleary, S. "The Relationship between Firm Investment and Financial Status." Journal of Finance, 54
(1999), 673-692.
Demirguc-Kunt, A., and V Maksimovic. "Law, Finance, and Firm Growth." Journal of Finance, 53
(1998), 2107-2137.
Demirguc-Kunt, A., and V. Maksimovic. "Funding Growth in Bank-Based and Market-Based Fi
nancial Systems: Evidence from Firm-Level Data." Journal of Financial Economics, 65 (2002),
337-363.
Erickson, T., and T. Whited. "Measurement Error and the Relationship between Investment and Q."
Journal of Political Economy, 108 (2000), 1027-1057.
Fairfield, P., and T. Yohn. "Using Asset Turnover and ProfitMargin to Forecast Changes in Profitabil
ity."Review of Accounting Studies, 6 (2001), 371-385.
Fazzari, S.; R. Hubbard; and B. Petersen. "Financing Constraints and Corporate Investment." Brook
ings Papers on Economic Activity, 1 (1988), 141-195.
Healy, P.; A. Hutton; and K. Palepu. "Stock Performance and Intermediation Changes Surrounding
Sustained Increases inDisclosure." Contemporary Accounting Research, 16 (1999), 485-520.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions
 380 Journal of Financial and Quantitative Analysis

Jensen,M., andW. Meckling. "Theory of the Firm: Managerial Behavior, Agency Costs and Owner
ship Structure." Journal of Financial Economics, 3 (1976), 305-360.
Judge, G.; W. Griffith; R. Hill; and T. Lee. The Theory and Practice of Econometrics. New York, NY:
JohnWiley (1980).
Kaplan, R., and L. Zingales. "Do Investment-Cash Flow Sensitivities Provide Useful Measures of
Financing Constraints?" Quarterly Journal of Economics, 112(1997), 169-215.
Korajczyk, R., and A. Levy. "Capital Structure Choice: Macroeconomic Conditions and Financial
Constraints." Journal of Financial Economics, 68 (2003), 75-109.
La Porta, R.; F. Lopez-de-Silanes; A. Shleifer; and R. Vishny. "Investor Protection and Corporate
Governance." Journal of Financial Economics, 58 (2000a), 3-27.
La Porta, R.; F. Lopez-de-Silanes; A. Shleifer; and R. Vishny. "Agency Problems and Dividend
Policies around theWorld." Journal of Finance, 55 (2000b), 1-33.
Lang, M. "Discussion of Stock Performance and Intermediation Changes Surrounding Sustained
Increases inDisclosure." Contemporary Accounting Research, 16 (1999), 521-534.
Lang, M., and R. Lundholm. "Cross-Sectional Determinants of Analyst Ratings of Corporate Disclo
sures." Journal of Accounting Research, 31 (1993), 246-271.
Lang, M., and R. Lundholm. "Corporate Disclosure Policy and Analyst Behavior." Accounting Re
view, 71 (1996), 467-492.
Lang, M., and R. Lundholm. "Voluntary Disclosure and Equity Offerings: Reducing Information
Asymmetry or Hyping the Stock?" Contemporary Accounting Research, 17 (2000), 623-662.
Leuz, C, and R. Verrecchia. "The Economic Consequences of Increased Disclosure." Journal of
Accounting Research, 38 (2000), 91-124.
Levine, R., and S. Zervos. "Stock Markets, Banks, and Economic Growth." American Economic
Review, 88 (1998), 537-558.
Lundholm, R., and L. Meyers. "Bring the Future Forward: The Effect of Disclosure on the Returns
Earnings Relation." Journal ofAccounting Research, 40 (2002), 809-839.
Myers, S. "Determinants of Corporate Borrowing." Journal of Financial Economics, 5 (1977), 147?
175.
Myers, S., and N. Majluf. "Corporate Financing and Investment Decisions When Firms Have Infor
mation that Investors Do Not Have." Journal of Financial Economics, 13 (1984), 187-221.
Nagar, V.; D. Nanda; and P.Wysocki. "Discretionary Disclosures and Stock-Based Incentives." Jour
nal of Accounting and Economics, 34 (2003), 383-309.
Neter, J.;M. Kutner; C. Nachtsheim; andW. Wasserman. Applied Linear Statistical Models. Chicago,
IL: Irwin (1996).
Poterba, J. "Comment on Financing Constraints and Corporate Investment." Brookings Papers on
Economic Activity, 1988, 200-204.
Rajan, R. "Insiders and Outsiders: The Choice between Informed and Arm's Length Debt." Journal
of Finance, M (1992), 1367-1400.
Rajan, R., and L. Zingales. "Financial Dependence and Growth." American Economic Review, 88
(1998), 559-586.
Sengupta, P. "Corporate Disclosure Quality and the Cost of Debt." Accounting Review, 73 (1998),
459_474.
Stein, J. "Agency Costs, Information, and Capital Investment." In Handbook of the Economics of
Finance, G. Constantinides, M. Harris, and R. Stulz, eds. Boston, MA: Elsevier/North-Holland
(2003).
Teoh, S. H.; I.Welch; and T. J.Wong. "Earnings Management and theLong-Run Market Performance
of InitialPublic Offerings." Journal of Finance, 53 (1998a), 1935-1974.
Teoh, S. H.; I.Welch; and T. J.Wong. "Earnings Management and theUnderperformance of Seasoned
Equity Offerings." Journal of Financial Economics, 50 (1998b), 63-99.
Verrecchia, R. "Discretionary Disclosure." Journal ofAccounting and Economics, 5 (1983), 179-194.
Verrecchia, R. "Essays on Disclosure." Journal of Accounting and Economics, 32 (2001), 97-180.
Welker, M. "Disclosure Policy, Information Asymmetry and Liquidity in Equity Markets." Contem
porary Accounting Research, 11 (1995), 801-827.
White, H. "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test forHet
eroskedasticity." Econometrica, 4 (1980), 817-838.
Yu, F. "Accounting Transparency and the Term Structure of Credit Spreads." Journal of Financial
Economics, 75 (2005), 223-242.

This content downloaded from 128.235.251.160 on Wed, 25 Mar 2015 02:03:06 UTC
All use subject to JSTOR Terms and Conditions