Your Governance or Mine?

Your Governance or Mine?

Andrew Ferguson, Matthew Grosse, Stephen Kean & Tom Scott

I n a series of industry reports, the resource sectorhas been rated as having the worst corporategovernance of any sector (WHK Horwath 2009).

The financial and popular media have cited these reportsin drawing attention to the issue: ‘Small shareholdersare being warned to read the fine print before theyinvest in small mining companies.’1 However, is singlingout the resource industry for criticism on corporategovernance valid? This paper empirically investigateswhether development stage entities (DSEs) in theAustralian resource sector have systematic differencesin corporate governance compared to similar sized firmsin other sectors. Following recent research, we argue thatsystemic differences in corporate governance are likelyexplained by the economic characteristics of resourceDSEs and suggest that the label of ‘bad’ governancemay not be appropriate (Armstrong et al. 2010; Brickleyand Zimmerman 2010).2 This result contributes to thewider corporate governance literature and is of interest toAustralian investors and regulators given the importanceof the junior resource sector to the Australian equitymarket in terms of aggregate constituent numbers.

Superior corporate governance is argued to beassociated with better firm performance (for example,Bhagat and Bolton 2008; Dahya et al. 2008), a lowercost of capital (for example, Anderson et al. 2004;Huang et al. 2009) and lower seasoned equity offeringunder-performance (Brown et al. 2009). Regulators alsoemphasise ‘good’ corporate governance, with the ASXissuing best practice guidelines. The financial pressincreasingly covers corporate governance, frequentlynaming the resource sector as the ‘worst offender’ basedon corporate governance scores.3 The WHK HorwathRating System awards firms up to five stars based on ASXCorporate Governance Council Principles. However,this methodology implicitly assumes that more cor-porate governance or greater compliance is ‘better’and rewards firms who ‘box-tick’, regardless of firmspecific or other institutional and economic factors. Thiscontrasts with a growing body of academic literature thatcriticises a ‘one size fits all’ or ‘more is better’ approachto assessing governance quality. Reviewing the literature,Armstrong et al. (2010) highlight the commonly heldmyth that it is easy to distinguish ‘good’ from ‘bad’ for

In response to criticism directed at the resource sector’scorporate governance, this paper examines the corporategovernance and underlying firm characteristics ofresource development stage entities (DSEs) relative to asize-matched sample of non-resource firms. We find thatresource DSEs have different governance characteristics inthe measures of board independence, chair/CEO dualityand CEO cash bonuses. Furthermore, there are differencesin the information environment measures of analystfollowing, debt levels, stock market return and stockturnover. Considering we document substantialdifferences in underlying firm characteristics, corporategovernance differences are likely appropriate to themining industry and should not be uniformly labelled as‘bad’. Our results suggest that media rankings based oncorporate governance scores may not accurately portraythe resource sector. Overall, our results are of interest toAustralian investors and regulators and contribute to abroader understanding of contextually contingentcorporate governance.

CorrespondenceAndrew Ferguson, Discipline of Accounting, UTS, PO Box123 Broadway NSW 2007, Australia. Tel: +61 2 9514 3565;fax: +61 2 9514 3669; email: [email protected]

doi: 10.1111/j.1835-2561.2011.00147.x

406 Australian Accounting Review No. 59 Vol. 21 Issue 4 2011

A. Ferguson, M. Grosse, S. Kean & T. Scott Your Governance or Mine?

any given corporate governance mechanism. This mythignores economic arguments and empirical evidence onwhy firms that are labelled as having bad governancemay have, for example, appropriately selected a boardwith low independence consistent with the firm specificsituation. Brickley and Zimmerman (2010) highlighthow this myth has extended into two further myths:(a) that it is possible to compute corporate governancescores; and (b) that it is possible to identify corporategovernance ‘best practice’.

Providing empirical evidence that one size does not fitall, Coles et al. (2008) find that research and development(R&D) intensive firms have larger boards and a higherproportion of executive directors. They interpret thisfinding as consistent with the board’s increased demandfor inside technical knowledge in order to properlyadvise and monitor management. Matolcsy and Wright(2007) document that equity-based compensation, anAmerican hallmark of ‘good’ corporate governance, is farless common amongst Australian firms. They interpretthis finding as consistent with the inappropriateness ofsuch compensation schemes for comparatively smallersized Australian firms. Patro et al. (2009) examine 82firms over 65 years and document that changes infirm size, growth opportunities, merger activity andgeographical expansion are associated with changes inboard characteristics. This suggests a wide range offactors affect governance and that a simple ‘good’ or ‘bad’label is misleading. Linden and Matolcsy (2004) find apositive association between WHK Horwath governanceratings and firm size in the largest 250 Australian firmsin 2001, illustrating the strong effect of firm size oncompliance. In summary, compliance with corporategovernance mechanisms is costly and firms must balancethe costs and benefits of specific mechanisms within theiroverall governance package (Armstrong et al. 2010).

Therefore, we argue that the underlying firmcharacteristics of resource DSEs help explain rationalcorporate governance choice, rather than a simple‘good’ or ‘bad’ dichotomy. As with R&D intensivefirms, resource DSEs are a firm type where corporategovernance mechanisms that may be classified as‘bad’ governance can be economically justified. Theorganisational characteristics of a typical resource DSEare summarised in a recent observation by TonyFeatherstone:

Boards of micro-cap explorers often seem out of stepwith modern governance practices. It is not uncommonfor a board of a newly listed explorer to have an executivechairman, managing director and only one independentdirector.4

The question we consider is whether the corporategovernance of resource DSEs is different from similarsized non-resource firms. Rather than singling out

resource explorers from the micro-cap end of the marketfor criticism, we investigate whether the corporategovernance differences that do exist between sizedmatched resource DSEs and non-resource firms areexplained by underlying firm characteristics. This canbe formally stated as:

H1: Differences between the corporate governancecharacteristics of resource and non-resource firms are afunction of firm characteristics.

Research Design

Sample selection

Following Ferguson et al. (2011), a firm is designatedas a resource DSE if its operating revenue is less than5% of market capitalisation. A preliminary sampleof 100 ASX-listed DSEs from the resource industries(GICS sectors 101020 and 151040, namely Oil, Gas andConsumable Fuels, and Metals and Mining, respectively)was randomly selected for the year 2009. To providea comparison, 100 non-resource firms were uniquelymatched on the basis of market capitalisation at 31st

December 2009. Three firms lacked a full 2009 annualreport, six firms were transitioning between resource andnon-resource industries, whilst two firms were undersuspension from active trading on the ASX. These firmswere replaced by a further 11 randomly selected firms forthe corresponding sub-sample. This left a final sample of100 resource DSEs and 100 size-matched non-resourcefirms.5 The year 2009 was chosen as it is a recent year withfull availability of annual reports. This study is limited toone year of data (2009) to avoid the econometric issuesinvolved with pooling multiple years of data when theunderlying corporate governance variables are sticky.6

The stickiness of corporate governance mechanisms isemphasised by Brown et al. (2011), who documentcorrelations of around 0.8 between adjacent years offirms’ Horwath corporate governance scores.

Due to the broad definition of corporate governancethere is a large range of potential measures to examine,for example Larcker et al. (2007) investigate 39 measures.Corporate governance studies frequently attempt tocondense various corporate governance mechanismsinto an objective corporate governance ‘score’. However,these scores may not facilitate a meaningful comparisonbetween strategies, life-cycle stages and operatingenvironments (Armstrong et al. 2010; Brickley andZimmerman 2010). Therefore, we examine a limitedrange of corporate governance variables that arecommonplace in the extant literature (Brown et al. 2011).They are not meant to be, nor could possibly be, anexhaustive list of all the possible mechanisms utilisedby corporations, however, they do provide a conciseoverview of popular governance measures.

C© 2011 CPA Australia Australian Accounting Review 407

Your Governance or Mine? A. Ferguson, M. Grosse, S. Kean & T. Scott

Table 1 Data definitions

Panel A: Corporate Governance Variables

ASX Principles Board Size Number of directors on the boardBoard Independence Percentage of self-designated independent directors

on the boardMajority Independence Frequency of boards with independence greater

than 50%Chair/CEO Duality Frequency of boards with the CEO as the chair

CEO Compensation CEO Cash Pay (’000) Total cash compensation for the CEOCEO Equity Plan Frequency of equity-based compensation plansCEO Cash Bonus Frequency of cash bonuses

Ownership Structure Substantial Shareholders Number of substantial shareholders (5% or more)Largest Shareholding Percentage of firm stock held by the largest

shareholderTop 20 Shareholdings Percentage of firm stock held by the top 20

shareholdersCEO Shareholdings Percentage of firm stock held by CEOBoard Shareholdings Percentage of firm stock held by Board (excl. CEO)KMP Shareholdings Percentage of firm stock held by KMP (excl.

executive directors)Audit Top-tier Frequency of top-tier audit firms (PWC, KPMG,

Deloitte and Ernst & Young)

Panel B: Firm Characteristics

Employees Key Management Personnel (KMP) Number of executives listed in the disclosure of top 5executive compensation

Employee Expense Total compensation paid to employees(excl. executive directors) divided by total assets

CEO/Employee Expense CEO cash pay divided by employee expensesFinancial Position Total Assets (’000) Total assets

Market-to-Book Market capitalisation divided by the book value ofordinary equity

Debt-to-Equity Debt divided by market capitalisation (restricted tofirms with debt)

New Debt Capital Purchase frequency of new debtLong-term Debt Capital Frequency of long-term debt (new and old)Current Ratio Current assets divided by current liabilities

Financial Performance Return-on-Assets Earnings before interest and tax divided by totalassets

Return-on-Equity Net Profit divided by the book value of ordinaryequity

Dividends Frequency of dividendsNegative Free Cash Flows Frequency of negative free cash flowsCash Burn Current assets minus current liabilities, divided by

free cash flows (restricted to firms with negativefree cash flows)

Stock Return Average monthly stock returnReturn Volatility Standard deviation of monthly stock return

Stock Market Activity Analyst Coverage Frequency of public equity analyst coverageNumber of Analysts Number of analysts (restricted to firms with analyst

coverage)Stock Turnover Average monthly stock volume divided by the

number of issued stockTurnover Volatility Standard deviation of monthly stock turnover

The variables we examine are based around thecomponents of the Horwath corporate governancerating methodology.7 First, we used measures ofboard size, board independence and CEO/board chairduality, which are also key ASX principles for goodcorporate governance. Measures of board subcommitteesize and independence were not examined due to

inapplicability and strong correlation to board sizeand independence when overall board size is small.Second, to allow quantitative analysis, we reportedCEO compensation, shareholding levels and shareholderconcentration rather than the existence of related policiesused in the Horwath corporate governance scores. Third,we substituted a measure of top-tier auditor for the

408 Australian Accounting Review C© 2011 CPA Australia


ratio of non-audit fees to audit fees. We argued thatinterpreting the audit fee ratio is complicated for smallfirms considering the practical reliance on outsourcingaccounting and consulting functions. Furthermore,top-tier auditor usage can potentially signal strongerexternal monitoring to the market (DeAngelo 1981). Tostructure our discussion, variables were classified intofour categories: ASX Principles for good governance;CEO compensation; ownership structure; and audit. Allcorporate governance data was manually collected fromannual reports and variables are defined in Table 1.

To allow the investigation of differences betweenresource and non-resource firms, we reported a rangeof underlying firm characteristics. They are classifiedinto four categories: employees; financial position;financial performance; and stock market activity. Stockmarket and analyst information was collected fromthe DataStream and I/B/E/S databases, respectively. Allother firm information came from FinAnalysis and wemanually collected missing data from annual reports.A full description of the underlying firm characteristicvariables accompanies Table 1.8

Methodology

To illustrate differences between resource and non-resource firms we report descriptive statistics on bothcorporate governance and firm characteristics. Wecompare whether the means of resource and non-resource firms for each variable are significantly differentusing Student t-tests. We also report whether the ranksof variables are significantly different using the Mann-Whitney U test and report z-statistics. This parametricand non-parametric univariate analysis allows us toidentify any trends or difference in the corporategovernance and firm characteristics of resource and non-resource firms. However, a potential caveat of univariatetests is that they only illustrate significant differences, notcausality or associated differences, leaving interpretationopen.

To provide further insight on whether there isa difference between the corporate governance ofresource and non-resource firms after controlling forunderlying firm characteristics, we conducted severalmultivariate tests. We regress corporate governancemeasures (CorpGov) on a categorical variable equalto one if the firm is in the resource sector (Resource)and firm characteristics (FirmChar). A significantcoefficient on Resource would suggest a difference inthe corporate governance of the resource sector thatis not explained by the firm characteristics included inthe regression. Firm characteristics are excluded wherethere is potential multi-collinearity or other econometricissues.9 We recognise that causal inferences, endogeneityand correlated omitted variable bias are particularly

problematic in corporate governance research (forexample, Armstrong et al. 2010; Brown et al. 2011).However, due to the exploratory nature of this paper andthe number of multivariate tests conducted, thoroughsensitivity testing and generalisability of specific resultsis a topic for future research. For continuous corporategovernance variables the ordinary least squares (OLS)regression is specified as:

CorpGov = α + β1Resourcei + β2FirmChari + ε

(1)

For categorical corporate governance variables thelogistic regression is specified as:

Logit (CorpGov = 1) = α + β1Resourcei

+β2FirmChari + ε (2)

Results

Univariate analysis of corporate governance

To identify any trends, we compared the means andmedians of corporate governance variables betweenresource DSEs and non-resource firms in Panel ATable 2. First, in contrast to media reports, resourceDSEs only have marginally smaller boards (means of4.3 to 4.6 directors). Second, the board independenceresults (means of 43% and 45% for resource and non-resource sub-samples, respectively) potentially highlightthe inapplicability of ASX recommendations andcorporate governance scoring systems to small firms.Third, both sectors typically separate the roles of CEOand chairman, but resource DSEs have higher rates ofchair/CEO duality than non-resource firms.

Mean cash pay for CEOs is similar, but the mediancash pay for non-resource CEOs is almost 50% more(z-stat = -3.018). We document that about half of bothtypes of CEOs (resource and non-resource) have equitycompensation plans, while Matolcsy and Wright (2007)document that about two-thirds of ASX 500 firms haveequity compensation in 1999–2001. This difference islikely a joint function of time and that our sample firmsare significantly smaller. However, resource DSE CEOshave a far lower cash bonus frequency (8% to 42%).Resource DSEs have nearly half the frequency of top-tierauditors as non-resource firms (28% to 46%).

Resource DSEs have lower mean ownership con-centration than non-resource firms (less substantialshareholders and smaller largest and top 20 share-holdings). The governance ramifications of ownershipstructure are difficult to interpret. Large independentshareholders have a strong economic incentive to



Tab

le2

Des

crip

tive

dat

a(2

009)

Pane

lA:C

orpo

rate

Gov

erna

nce

Varia

bles

Reso

urce

DSE

sN

on-R

esou

rce

Firm

sC

ompa

rison

s

Mea

nSt

d.D

ev.

Min

Med

ian

Max

Mea

nSt

d.D

ev.

Min

Med

ian

Max

t-st

atz-

stat

ASX

Prin

cipl

esBo

ard

Size

4.3

1.6

3.0

4.0

10.0

4.6

1.3

3.0

5.0

9.0

−2.0

57∗∗

−2.6

87∗∗

∗

Boar

dIn

depe

nden

ce43

%25

%0%

38%

100%

45%

24%

0%50

%10

0%−0

.739

−0.6

61

Maj

ority

Inde

pend

ence

33%

38%

Cha

ir/C

EOD

ualit

y19

%8%

CEO

Com

pens

atio

nC

EOC

ash

Pay

(’000

)$5

13.4

0$8

13.1

1$2

1.80

$262

.66

$5,3

45.0

0$5

71.4

3$6

45.7

6$2

2.73

$381

.98

$4,0

39.3

8−0

.829

−3.0

18∗∗

∗

CEO

Equi

tyPl

an49

%49

%C

EOC

ash

Bonu

s8%

42%

Ow

ners

hip

Stru

ctur

eSu

bsta

ntia

lSh

areh

olde

rs2.

41.

40.

02.

06.

03.

41.

70.

03.

08.

0−4

.312

∗∗∗

−4.0

52∗∗

∗

Larg

est

Shar

ehol

ding

17%

17%

0%13

%82

%25

%19

%4%

18%

99%

−3.0

22∗∗

∗−3

.400

∗∗∗

Top

20Sh

areh

oldi

ngs

58%

17%

11%

60%

98%

69%

16%

25%

70%

100%

−4.7

29∗∗

∗−4

.636

∗∗∗

CEO

Shar

ehol

ding

s5%

9%0%

3%62

%9%

15%

0%2%

88%

−1.8

90∗

0.40

10Bo

ard

Shar

ehol

ding

s7%

10%

0%3%

47%

14%

15%

0%9%

71%

−3.6

03∗∗

∗−2

.893

∗∗∗

KM

PSh

areh

oldi

ngs

0%1%

0%0%

5%1%

2%0%

0%13

%−3

.205

∗∗∗

−3.3

52∗∗

∗

Aud

itTo

p−tie

r28

%46

%

Pane

lB:F

irmC

hara

cter

istic

s

Empl

oyee

sK

MP

3.1

1.4

1.0

3.0

5.0

4.2

1.3

0.0

5.0

5.0

−5.7

96∗∗

∗−5

.725

∗∗∗

Empl

oyee

Expe

nse

10%

21%

0%4%

149%

38%

59%

0%24

%52

2%−4

.292

∗∗∗

−7.3

23∗∗

∗

CEO

/Em

ploy

eeEx

pens

e72

9%60

90%

1%52

%60

976%

47%

182%

0%7%

1479

%1.

120

8.49

1∗∗∗

Fina

ncia

lPos

ition

Tota

lAss

ets

(’000

)$6

2.07

$196

.51

$0.2

7$1

1.23

$1,8

03.6

7$1

86.8

3$5

69.3

8$1

.02

$31.

80$4

,247

.18

−2.9

86∗∗

∗−3

.325

∗∗∗

Mar

ket-

to-B

ook

5.23

12.0

1−3

4.46

2.08

91.5

29.

5067

.17

−17.

181.

7667

3.28

−0.6

201.

747∗

Deb

t-to

-Equ

ity0.

011

0.04

70.

000

0.00

00.

271

0.27

70.

731

0.00

00.

017

5.20

7−3

.560

∗∗∗

−2.4

53∗∗

New

Deb

tC

apita

l17

%54

%Lo

ng-t

erm

Deb

tC

apita

l20

%60

%

Cur

rent

Ratio

14.9

118

.59

0.18

7.76

94.8

13.

477.

490.

131.

5659

.64

5.63

8∗∗∗

6.95

6∗∗∗

Con

tinue

d



Tab

le2

(Co

nti

nu

ed)

Pane

lB:F

irmC

hara

cter

istic

s

Reso

urce

DSE

sN

on-R

esou

rce

Firm

sC

ompa

rison

s

Mea

nSt

d.D

ev.

Min

Med

ian

Max

Mea

nSt

d.D

ev.

Min

Med

ian

Max

t-s

tat

z-st

at

Fina

ncia

lRe

turn

-on-

Ass

ets

−74%

145%

−732

%−2

1%1%

−20%

61%

−262

%3%

50%

−3.4

59∗∗

∗−6

.465

∗∗∗

Perf

orm

ance

Retu

rn-o

n-Eq

uity

−86%

234%

−120

3%−2

4%72

0%−1

97%

1783

%−1

7732

%0%

1623

%0.

613

−4.2

51∗∗

∗

Div

iden

ds0%

27%

Neg

ativ

eFr

eeC

ash

Flow

s93

%57

%

Cas

hBu

rn12

.980

111.

526

0.01

10.

792

1076

.770

2.33

16.

917

0.00

00.

592

51.0

900.

955

0.76

1St

ock

Retu

rn7%

6%−5

%7%

31%

2%6%

−18%

1%21

%5.

483∗

∗∗4.

970∗

∗∗

Retu

rnVo

latil

ity26

%11

%0%

25%

57%

22%

13%

0%20

%69

%2.

356∗

∗2.

734∗

∗∗

Stoc

kM

arke

tA

naly

stC

over

age

16%

38%

Act

ivity

Num

ber

ofA

naly

sts

2.62

52.

553

1.00

02.

000

11.0

004.

526

4.61

31.

000

2.00

016

.000

−3.3

05∗∗

∗−1

.000

Stoc

kTu

rnov

er49

.22

58.1

00.

0033

.25

412.

4429

.51

44.2

70.

0010

.43

286.

502.

877∗

∗∗4.

194∗

∗∗

Turn

over

Vola

tility

48.9

570

.56

0.00

24.8

146

4.47

30.9

980

.78

0.00

9.97

746.

921.

755∗

3.92

0∗∗∗

Tabl

e2

pres

ents

desc

riptiv

est

atis

tics

and

univ

aria

tean

alys

is(p

aram

etric

and

non-

para

met

ric)

for

corp

orat

ego

vern

ance

and

firm

char

acte

ristic

varia

bles

.Sa

mpl

eco

nsis

tsof

100

reso

urce

DSE

san

d10

0no

n-re

sour

cefir

ms

mat

ched

onm

arke

tca

pita

lisat

ion.

All

varia

bles

defin

edin

Tabl

e1.

p-va

lues

are

mat

ched

and

two-

taile

d.∗∗

∗p

<0.

01,∗

∗p

<0.

05,∗

p<

0.10

.



monitor management (Shleifer and Vishny 1986) andmanagement shareholdings bond their interests toshareholders (Jensen and Meckling 1976). However,particularly large shareholders can act against theinterests of the minority shareholders (Morck et al.1988; McConnell and Servaes 1990; Shleifer and Vishny1986). Although ownership concentration does have aneffect on corporate governance, it is not a specificallydesigned corporate governance mechanism and isinfluenced by other factors. Therefore, interpreting thedeterminants of ownership concentration through acorporate governance framework is highly problematic(Armstrong et al. 2010).

Resource firms also have significantly lower board(excluding CEO) and key management personnel(KMP) shareholdings (excluding executive directors).This smaller shareholding could be due to the earlylife cycle stage of resource DSEs. The lower KMPshareholding is likely a mechanical function of the fewerKMP in resource DSEs as outlined below, although italso suggests that organisational size is a key driverof total KMP shareholding. However, there are otherpotential explanations for these results. We do notreport regression results for ownership concentrationmeasures due to the lack of theoretical grounding, whichmakes interpretation of any significant determinantsproblematic.

Univariate analysis of firm characteristics

To inform our analysis on whether any observedcorporate governance differences can be explainedby underlying factors, we next discuss the firmcharacteristics of resource and non-resource firms asoutlined in Panel B Table 2. First, resource DSEs have asignificantly lower number of KMP disclosed in annualreports, a smaller non-executive employee expense anda higher CEO pay to employee expense ratio (mean of729% relative to 47%).10 These results support anecdotalevidence that resource DSEs have as few as one ortwo executive directors who are normally involved inproject acquisition or project generation activities, suchas designing drilling programs for existing projects.These executives are often accompanied on the board bynon-executives who routinely hold significant industryexperience and act in an advisory capacity. ResourceDSEs may also employ a receptionist as the only otheremployee, with other tasks outsourced.

Second, resource DSEs have significantly less longterm debt and leverage, with a mean debt-to-equityratio of 0.011 for resource DSEs relative to 0.277for non-resource firms. Furthermore, resource DSEshave a higher proportion with negative free cash flows(93% for resource, 57% for non-resource).11 Consistentwith capital scarcity and cash conservation for mineral

exploration, resource DSEs have higher current ratios,reflecting their need to raise cash to fund exploration,but similar rates of cash burn (amongst the firmswith negative free cash flows). This result is similar toresource DSE descriptive evidence reported in Ferguson(2011) and suggests that debt-based monitoring is veryrare and that resource exploration is high risk. Asthe non-resource firms are matched based on marketcapitalisation, the significantly lower total assets ofresource DSEs reflects the lack of debt financing.This is an intuitive result, as typically the only realassets on resource DSE balance sheets are cash anddeferred acquisition and exploration expenditure (whichis subject to impairment provisions under AASB 6).Resource DSEs also have substantially lower profitability.Therefore, as traditional profitability analysis cannotbe performed on resource DSEs, non-GAAP drilling,resource and reserve disclosure likely takes the placeof GAAP financial performance-related information(Ferguson and Crocket 2003).

Third, our univariate results suggest that resourceDSEs have higher stock volatility than non-resourcefirms consistent with Ball and Brown (1980). Fur-thermore, over this constrained sample window, sharereturns for resource DSEs are higher than for non-resource firms. This is likely to be primarily a functionof the commodity boom but may also be a partialreflection of a small number of distressed firms in thenon-resource sample. Share turnover, which can proxyfor stock liquidity, is also higher for resource DSEs.12

Fourth, non-resource firms have twice the frequencyof analyst following and almost twice the number ofanalysts (amongst the followed firms). This observationsuggests that despite the potentially higher demand foranalysts as information intermediaries and a possiblemonitoring mechanism, analysts are more reluctant topublicly follow resource DSEs. Although we do notempirically examine this difference, this equity analystreluctance could be due to a lack of technical skills(analysts typically have a finance-based education, notgeological or engineering-based).13

Overall, on a descriptive level our results suggestthat resource DSEs have fundamentally differentfirm characteristics from non-resource firms. Broadlyspeaking, they have less long term debt, more cashon hand, better market performance and lower analystcoverage.

Multivariate analysis

Results from OLS regressions of the continuouscorporate governance variables are reported in Table 3,with both models reporting F-statistics significant at the1% level and adjusted R2 above 40%. Furthermore, theVariance Inflation Factors for the OLS Models I and II



Table 3 Multivariate analysis of continuous corporategovernance variables

Model I Model IIBoard Size CEO Cash Pay ($m)

Coeff. p-value Coeff. p -value

Constant 3.040 0.000∗∗∗ 0.196 0.020∗∗

Resource −0.124 0.587 −0.027 0.610KMP 0.115 0.084∗ 0.011 0.453Employee

Expense0.355 0.360 0.174 0.053∗

Total Assets($m)

0.007 0.000∗∗∗ 0.001 0.000∗∗∗

Market-to-Book

0.025 0.150 0.001 0.867

Long-termDebt Capital

−0.174 0.361 −0.052 0.234

Return-on-Assets

0.252 0.113 0.014 0.704

Negative FreeCash Flow

0.528 0.033∗∗ −0.018 0.751

Cash Burn 0.074 0.307 0.011 0.500Stock Return 0.719 0.636 0.696 0.048∗∗

Return Volatility 0.029 0.971 0.110 0.544Analyst

Coverage0.419 0.091∗ 0.121 0.034∗∗

Stock Turnover 0.006 0.265 0.004 0.000∗∗∗

TurnoverVolatility

−0.007 0.166 −0.004 0.000∗∗∗

Adjusted R2 0.401 0.474F-statistic 10.502 0.000∗∗∗ 13.794 0.000∗∗∗

Table 3 presents the results of OLS regressions for a selectionof continuous corporate governance variables. Sample consistsof 100 resource DSEs and 100 non-resource firms matched onmarket capitalisation. All variables defined in Table 1. p-values arematched and two-tailed. ∗∗∗ p < 0.01, ∗∗ p < 0.05, ∗ p < 0.10.

reported in Table 3 are lower than 6.8, suggesting noproblematic multi-collinearity (Lardaro 1993). Resultsfrom logistic regressions of the categorical corporategovernance variables are reported in Table 4, with allmodels reporting Chi squares significant at the 1% leveland Pseudo R2 ranging from 20% in Model I (MajorityBoard Independence) to 47% in Model IV (CEO CashBonus).

ASX principles

OLS regression results from Table 3 Model I providefurther confidence that there is no significant differencebetween the board size of resource and non-resourcefirms after controlling for differences in firm character-istics. This is contrary to evidence of larger boards insectors where highly technical idiosyncratic informationmight be in greater demand, but consistent with thelower operational diversity of resource DSEs, who,given cash constraints often focus on the developmentof a single project (Coles et al. 2008). Total assetsand analyst coverage are positively associated with Tab

le4

Mu

ltiv

aria

tean

alys

iso

fca

teg

ori

calc

orp

ora

teg

ove

rnan

ceva

riab

les

Mod

elI

Mod

elII

Mod

elIII

Mod

elIV

Mod

elV

Maj

ority

Inde

pend

ence

Cha

ir/C

EOD

ualit

yC

EOEq

uity

Plan

CEO

Cas

hBo

nus

Top-

tier

Aud

itor

Coe

ff.

p-v

alue

Coe

ff.

p-v

alue

Coe

ff.

p-va

lue

Coe

ff.

p-va

lue

Coe

ff.

p-v

alue

Con

stan

t0.

837

0.26

5−1

.216

0.27

2−2

.144

0.00

6∗∗∗

−0.9

380.

354

−1.2

880.

137

Reso

urce

−0.8

860.

067∗

1.51

90.

050∗∗

0.37

00.

445

−1.1

520.

071∗

0.32

00.

565

KM

P−0

.187

0.17

8−0

.043

0.82

30.

286

0.03

7∗∗0.

176

0.37

9−0

.078

0.62

8Em

ploy

eeEx

pens

e−1

.201

0.15

5−1

.740

0.20

71.

655

0.04

0∗∗−0

.331

0.74

03.

374

0.00

1∗∗∗

Tota

lAss

ets

($m

)0.

000

0.99

30.

002

0.60

2−0

.003

0.25

1−0

.001

0.84

90.

010

0.00

6∗∗∗

Mar

ket-

to-B

ook

0.03

30.

362

0.03

80.

434

0.02

70.

463

0.02

60.

621

−0.0

120.

770

Long

-ter

mD

ebt

Cap

ital

0.12

60.

755

1.47

10.

015∗∗

−0.1

890.

633

0.61

60.

215

−0.1

890.

685

Retu

rn-o

n-A

sset

s−0

.679

0.04

6∗∗−0

.790

0.07

8∗0.

214

0.51

90.

390

0.46

10.

639

0.14

9N

egat

ive

Free

Cas

hFl

ow−0

.599

0.24

2−0

.564

0.44

30.

886

0.09

2∗−1

.768

0.00

3∗∗∗

−0.6

620.

270

Cas

hBu

rn−0

.054

0.72

8−0

.235

0.39

00.

039

0.79

50.

277

0.16

10.

262

0.11

0St

ock

Retu

rn3.

427

0.29

74.

210

0.39

93.

626

0.25

7−3

.500

0.42

65.

296

0.16

3Re

turn

Vola

tility

−4.2

450.

018∗∗

−6.1

750.

027∗∗

−1.9

300.

237

−1.4

650.

514

−1.8

000.

358

Ana

lyst

Cov

erag

e0.

308

0.55

1−1

.272

0.19

30.

991

0.06

1∗1.

636

0.00

6∗∗∗

1.05

10.

057∗

Stoc

kTu

rnov

er0.

021

0.05

1∗−0

.026

0.19

00.

027

0.02

2∗∗0.

007

0.59

0−0

.006

0.63

4Tu

rnov

erVo

latil

ity−0

.002

0.82

50.

016

0.37

1−0

.027

0.01

8∗∗−0

.003

0.81

10.

006

0.62

4N

agel

kerk

eR2

0.20

70.

255

0.22

60.

475

0.44

1C

hisq

uare

32.6

750.

003∗∗

∗30

.069

0.00

7∗∗∗

37.1

100.

001∗∗

∗77

.373

0.00

0∗∗∗

77.9

680.

000∗∗

∗

Tabl

e4

pres

ents

the

resu

ltsof

logi

stic

regr

essi

ons

for

ase

lect

ion

ofca

tego

rical

corp

orat

ego

vern

ance

varia

bles

.Sam

ple

cons

ists

of10

0re

sour

ceD

SEs

and

100

non-

reso

urce

firm

sm

atch

edon

mar

ket

capi

talis

atio

n.A

llva

riabl

esde

fined

inTa

ble

1.p-

valu

esar

em

atch

edan

dtw

o-ta

iled.

∗∗∗

p<

0.01

,∗∗

p<

0.05

,∗p

<0.

10.



board size, suggesting a strong correlation betweenfirm size or project development and board size.Alternatively, if governance is linked to performance,analysts would rationally choose to follow firms with‘better’ governance practices. We interpret the positivecoefficients on KMP and negative free cash flow as con-sistent with board representation of management andcreditors.

Table 4 Model I presents a logistic regressionfor boards with independence over/under 50%. Thesignificant negative coefficient on Resource (p = 0.067),suggests that after controlling for firm characteristics,resource DSEs are less likely to have a majority ofindependent directors. This result departs from the lackof a significant univariate difference in the percentageof independent directors outlined in Table 2, butis consistent with the greater proportion of insiderdirectorship documented in other sectors where highlytechnical knowledge and experience is likely to be ingreater demand (Coles et al. 2008). We also documentthat lower return volatility and higher turnoverare positively associated with board independence,suggesting that the market either enforces, or is attractedto, more independent boards. The negative associationbetween return on assets (ROA) and board independenceis likely a joint function of the large number ofloss making companies in our sample as discussedabove and because poorly performing firms are morelikely to increase board independence (Bhagat andBlack 2002).

The difference in the duality of resource firmsdocumented by the univariate results is supported bya positive Resource coefficient in respect to chair/CEOduality in Table 4 Model II. This could reflect theimportance of technical knowledge in guiding boarddiscussions (Brickley et al. 1999). However, we cannotconclude this with any certainty and there are otherpossible explanations, such as more powerful or skilfulCEOs bargaining for the chairman role (Hermalinand Weisbach 1998). Furthermore, long-term debtand return volatility are positively associated withchair/CEO duality. This could be due to the large numberof firms in our resource sub-sample without debtfinancing. Alternatively, it could suggest benefits fromconcentrated control or long-term relationships. Thenegative association between ROA and duality contrastswith literature that poorly performing firms are morelikely to separate the role of CEO and chairman (Chenet al. 2008). The different result is likely due to the greaterrelevance of non-financial information as a performancemeasure in the resource sector (Ferguson et al. 2011).

In summary, our results support our hypothesis bysuggesting that resource DSEs have greater duality and alower frequency of independent boards, consistent withother sectors where highly technical knowledge is ofgreater importance.

CEO compensation

In contrast to the univariate results, Table 3 Model IIfinds that Resource is not significantly associated withCEO cash pay, suggesting that the difference is a functionof firm characteristics. The regression results indicatethat the CEOs of more complex, high performing andriskier firms receive higher pay. As discussed above,resource DSEs have higher market returns, turnoverand volatility, consistent with CEO pay being a factor ofrationally risk adjusted compensation packages (Conyonet al. 2011).

The logistic regression results in Table 4 Model IIIshow that there is no significant difference in equityplan usage by resource DSEs and that equity planusage is predominantly a function of firm characteristicsthat arguably represent size, risk and performance.As discussed above, resource DSEs had higher stockreturn, turnover and volatility in 2009. This suggeststhat equity plan usage is appropriate for small resourceDSEs, considering that stock market efficiency is a keyrequirement for equity-based incentives to be an effectivebonding mechanism.14

Consistent with univariate results, Resource issignificantly negatively associated to CEO cash bonusin our logistic specification in Table 4 Model IV. Thenegative association between CEO cash bonus andnegative free cash flow suggests that discretionary cashcompensation for CEOs is highly dependent on firm cashposition. Therefore, lower cash bonus use by resourceDSEs is unsurprising given the significantly greater cashconstraints outlined above and is consistent with afurther quote from Featherstone who suggests:

Speculative explorers usually do not have the funds –or the need – for large boards with diverse skills. Nordo they have the cash for high director fees or executivesalaries.15

The positive association between analyst coverage andall aspects of CEO compensation could suggest thatanalysts are more likely to cover firms with ‘star’ CEOsor that analyst coverage is a proxy for future cash flowexpectations or size effects. Overall, our results supportour hypothesis by suggesting that resource DSEs onlyhave different CEO compensation in terms of cash bonususe, which is consistent with the greater cash constraintsfaced by resource DSEs.

Audit

After controlling for firm characteristics, the logisticregression results in Table 4 Model V show that Resourceis not significantly associated with lower top-tier auditoruse. This contrasts with the univariate result in Table2 and the significantly positive association between



top-tier auditor use and employee expense, total assetsand analyst coverage suggests that operational size orcomplexity is a key determinant of auditor choice.Furthermore, lower top-tier auditor use by firms withless total assets, but similar market capitalisation, isconsistent with a lower demand for premium auditservices amongst firms with a higher proportion of off-balance sheet (unaudited) assets (Anderson et al. 1993).Additionally, the greater value of non-GAAP drilling,resource and reserve disclosure for the resource sector(Ferguson et al. 2011), reduces the value of expensiveGAAP assurance. Lower top-tier auditor use is alsoconsistent with the cash constraints of resource DSEs(Lee et al. 2003), discussed above. Therefore, we arguethe difference in top-tier auditor use by resource DSEsis explained by firm characteristics, consistent with ourhypothesis.

Conclusion

Large accounting scandals (such as Enron, Parmalat andHIH Insurance), as well as the more recent issues with thetoxic legacy assets associated with the Global FinancialCrisis, have encouraged the financial press to nameand shame public companies with perceived corporategovernance deficiencies. Despite academic researchstressing the complexities surrounding the demand andsupply of individual mechanisms within a corporategovernance package, corporate governance scoringsystems have become popular in the financial press. Thispaper aims to consider whether the governance choicesmade by resource DSEs are economically justifiable andreflect fundamental differences in the way industriesfunction. In doing so, we provide a deeper level ofanalysis of the differences in corporate governancebetween resource and non-resource firms, as systematicdifferences are likely associated with underlying firmcharacteristics as opposed to exclusively non-complianceor mismanagement.

We document substantial differences in firm charac-teristics. First, resource DSEs typically have a smallerorganisational structure as illustrated by fewer manage-ment personnel and employees.16 Second, resource DSEshave less debt-based monitoring, consistent with debtcapital being expensive for firms with high operatingrisk. Third, higher stock market participation forresource DSEs relative to non-resource firms, warrantsdeeper analysis controlling for the effects of the cyclicalnature of the commodity market. Fourth, the loweranalyst coverage of resource DSEs, potentially suggests aless developed information environment.

After controlling for firm characteristics, popularmetrics, such as board size and top-tier auditor use aresimilar between the two sub-samples. However, resourceDSEs exhibit a higher frequency of chair/CEO dualityand a lower frequency of majority board independence.

As suggested, this is consistent with the greater value oftechnical competence in the resource sector. ResourceDSEs also have lower cash bonus usage for their CEOscompared to non-resource firms, but similar usage ofequity compensation, consistent with cash constraints.Univariate results depict resource firms having lowerownership concentration but results are difficult tointerpret.

Overall, we demonstrate that most differences ingovernance characteristics between resource DSEs andsimilar sized non-resource firms are a function of differ-ences in underlying firm characteristics and are unlikelyto be caused by management intentionally choosingpoor governance or non-compliance. Furthermore, ourevidence suggests that media criticisms of the resourcesector are likely exaggerated and equally applicable tosmall firms more generally. If poor governance wasendemic in the Australian resource industry, the reportedstrong share price performance of the resource sectorsample in this study and more broadly could not bereconciled in an efficient capital market. In light ofsuch evidence, we suggest that consistent with Brickleyand Zimmerman (2010), governance scoring systemsand ‘box-ticking’ exercises may in fact be misleading.17

The media and regulators alike should consider morecarefully the implications of fundamental differencesin the economics underpinning governance choices indifferent industries. Our analysis is exploratory andwe encourage future research that rigorously examinesindividual corporate governance mechanisms.

Andrew Ferguson, Matthew Grosse, Stephen Kean and TomScott are at University of Technology, Sydney.

Notes

1 Glanville, B. 2007, ‘Corporate governance being forgotten:report’, ABC News, 30 November 2007, available at www.abc.net.au/news/stories/2007/11/30/2105730.htm last accessed 5 June2011.

2 This study relies on the theoretical framework of previouscorporate governance literature. It is not the objective of thispaper to posit any alterative framework or empirical modelfor explaining corporate governance per se or to examine theperformance ramifications of corporate governance systems indifferent contexts.

3 WHK Horwath ‘2009 Corporate Governance Report: Companies’,p. 3.

4 Featherstone, T. 2010, ‘A new gold rush’, Company DirectorMagazine, September.

5 We classified 652 ASX listed resource firms as a DSE in 2009.Due to the problematic nature of matching at the smaller endof the market, there are 347 unique firm-pair matches betweenresource DSEs and non-resource firms. Due to the lack of machinereadable data, issues with delisting, market suspensions and noannual reports and the laborious nature of data collection, welimit ourselves to 200 observations.

6 Future research may wish to consider whether the evidence (andinferences) in this paper are subject to variation across time.Corporate governance trends are influenced by regulatory trends



(Tyler et al. 2010). However, the association between underlyingfirm characteristics and corporate governance is unlikely to varysignificantly from year to year.

7 As outlined in the ‘2009 WHK Horwath Corporate GovernanceReport’, (pp. 6–9).

8 A potential methodological issue with doing research on smallfirms is poor database coverage. Furthermore, small firms canhave legitimate observations that are well outside the expectednorm due to changing conditions or firm structure.

9 For example, number of analysts and analyst coverage are highlycorrelated; therefore we exclude number of analysts. Results arequalitatively similar to alternate financial characteristics inclusionand exclusion.

10 Eleven Resource DSEs and four non-resource firms did notreport a separate employee expense. For these firms the nearestequivalent (typically administrative expense) was used instead,results are robust to their exclusion.

11 Of the 57 non-resource firms, 12 were DSEs (primarily from thepharmaceutical development industry), whilst a further 28 wereloss-making firms and a further six were financially distressed(Altman 1968).

12 Once again it is difficult to disentangle the cyclical effects ofthe commodity boom from the underlying share turnover. In thisregard, future research may wish to consider a wider time frame tocontrol for cyclical effects (Ball and Brown, 1980). Future analysismay also wish to disentangle the effects of non-resource DSEs andfinancially distressed firms.

13 However, caution needs to be exercised in interpreting this analystresult given the dominance of boutique analyst resource houseswhich can be excluded from the I/B/E/S measure employed here.

14 We acknowledge the difficulty in, and importance of, disentan-gling the effect of the commodity cycle on stock price performancefrom managerial influences when designing an effective equity-based compensation system.

15 Featherstone, T. 2010, ‘A new gold rush’, Company DirectorMagazine, September.

16 Fewer employees is also consistent with the existence of cashconstraints, and future research matching sub-samples on thebasis of number of employees may suggest that resource DSEshave ‘better’ corporate governance than non-resource firms.

17 Although the comparison of corporate governance scores can bemeaningful if the underlying firm characteristics are identical.However, commentators have questioned whether it is possiblefor firm characteristics to ever be identical, whether corporategovernance and firm characteristics can be exhaustively measuredand efficient contracting would suggest that residual differencesare transitory.

References

Altman, E. 1968, ‘Financial Ratios, Discriminant Analysis andthe Prediction of Corporate Bankruptcy’, Journal of Finance,23, 4: 589–609.

Anderson, D., Francis J. and Stokes, D. 1993, ‘Auditing,Directorships and the Demand for Monitoring’, Journal ofAccounting and Public Policy, 12, 4: 353–75.

Anderson, R., Mansi, S. and Reeb, D. 2004, ‘BoardCharacteristics, Accounting Report Integrity, and the Costof Debt, Journal of Accounting and Economics, 37, 3:315–342.

Armstrong, C., Guay, W. and Weber, J. 2010, ‘The RoleOf Information and Financial Reporting in Corporate

Governance and Debt Contracting’, Journal of Accounting andEconomics, 50, 2–3: 179–234.

Ball, R. and Brown, P. 1980, ‘Risk and Return from EquityInvestments in the Australian Mining Industry: January1958 – February 1979’, Australian Journal of Management , 5, 1:45–66.

Bhagat, S. and Black, B. 2002, ‘The Non-correlation BetweenBoard Independence and Long-Term Firm Performance’,Journal of Corporation Law, 27, 2: 231–73.

Bhagat, S. and Bolton, B. 2008, ‘Corporate Governance andFirm Performance’, Journal of Corporate Finance, 14, 3: 257–73.

Brickley, J., Coles, J. and Linck, J. 1999. ‘What Happens ToCEOs After They Retire? New Evidence on Career Concerns,Horizon Problems, and CEO Incentives’, Journal of FinancialEconomics, 52, 3: 341–77.

Brickley, J. and Zimmerman, J. 2010, ‘Corporate GovernanceMyths: Comments on Armstrong, Guay, and Weber’, Journal ofAccounting and Economics, 50, 2–3: 179–234.

Brown, P., Lee, M. and Walter, T. 2009, ‘Corporate Governanceand the Long-Run Performance of Firms Issuing SeasonedEquity: An Australian Study’, 22ndAustralasian Finance andBanking Conference 2009.

Brown, P., Beekes, W. and Verhoeven, P. 2011, ‘CorporateGovernance, Accounting and Finance: A Review’, Accountingand Finance, 51, 1: 96–172.

Chen, C., Lin, J. and Yi, B. 2008, ‘CEO Duality and FirmPerformance: An Endogenous Issue’, Corporate Ownership andControl, 6, 1: 58–65.

Coles, J., Daniel, N. and Naveen, L. 2008, ‘Boards: DoesOne Size Fit All?’, Journal of Financial Economics, 87, 2:329–56.

Conyon, M., Core, J. and Guay, W. 2011, ‘Are US CEOs PaidMore Than UK CEOs? Inferences From Risk-Adjusted Pay?’,Review of Financial Studies, 24, 2: 402–38.

Dahya, J., Dimitrov, O. and McConnell, J. 2008, ‘DominantShareholders, Corporate Boards and Corporate Value: A Cross-Country Analysis’, Journal of Financial Economics, 87, 1: 73–100.

DeAngelo, L. 1981, ‘Auditor Size and Audit Quality’, Journal ofAccounting and Economics, 3, 3: 183–99.

Featherstone, T. 2010, ‘A New Gold Rush’, Company DirectorMagazine, September.

Ferguson, A. and Crockett, A. 2003, ‘Information Transfer andPress Coverage: The Case of the Gawler Craton Gold Boom’,Pacific-Basin Finance Journal, 11, 1: 101–20.

Ferguson, A., Clinch, G. and Kean, S. 2011, ‘PredictingFailure of Development Mining Projects’, Australian AccountingReview, 21, 1: 44–53.

Ferguson, A. 2011, Non-GAAP Voluntary Disclosure in anUnregulated Financial Statement Void: Determinants andImplications. Working Paper, UTS.



Glanville, B. 2007, ‘Corporate Governance Being Forgot-ten: Report’, ABC News, 30 November 2007, availableat www.abc.net.au/news/stories/2007/11/30/2105730.htm, lastaccessed 5 June 2011.

Hermalin, B. and Weisbach, M. 1998, ‘Endogenously ChosenBoards of Directors and Their Monitoring of the CEO’,American Economic Review, 88, 1: 96–118.

Huang, H., Wang, Q. and Zhang, X. 2009, ‘The Effect of CEOOwnership and Shareholder Rights on Cost of Equity Capital’,Corporate Governance, 9, 3: 255–70.

Jensen, M. and Meckling, W. 1976, ‘Theory of theFirm: Managerial Behaviour, Agency Costs and OwnershipStructure’, Journal of Financial Economics, 3, 4: 305–60.

Larcker, D., Richardson, S. and Tuna, I. 2007,‘Corporate Governance, Accounting Outcomes, andOrganizational Performance’, Accounting Review, 82, 4: 963–1008.

Lardaro, L. 1993, Applied Econometrics, HarperCollins, NewYork.

Lee, P., Stokes, D., Taylor, S. and Walter, T. 2003, ‘TheAssociation Between Audit Quality, Accounting Disclosuresand Firm-Specific Risk: Evidence From Initial Public Offerings’,Journal of Accounting and Public Policy, 22, 5: 377–400.

Linden, P. and Matolcsy, Z. 2004, ‘Corporate GovernanceScoring Systems: What Do They Tell Us?’, Australian AccountingReview, 14, 1: 9–16.

Matolcsy, Z. and Wright, A. 2007, ‘Australian CEO Compensa-tion: The Descriptive Evidence’, Australian Accounting Review,17, 3: 47–59.

McConnell, J. and Servaes, H. 1990, ‘Additional Evidence onEquity Ownership and Corporate Value’, Journal of FinancialEconomics, 27, 2: 595–612.

Morck, R., Shleifer, A. and Vishny, R. 1988, ‘ManagementOwnership and Market Valuation’, Journal of FinancialEconomics, 20, 1: 293–315.

Patro, S., Lehn, K. and Zhao, M. 2009, ‘Determinants of theSize and Structure of Corporate Boards: 1935–2000’, FinancialManagement , 38, 4: 747–80.

Shleifer, A. and Vishny, R. 1986, ‘Large Shareholders andCorporate Control’, Journal of Political Economy, 94, 3: 461–88.

Tyler, J., Matolcsy, Z. and Wells, P. 2010, ‘Was CorporateGovernance Regulation Really the Answer?’, UTS Workingpaper.

WHK Horwath. 2009, ‘2009 Corporate Governance Report’.


Date post:	07-Nov-2016
Category:	Documents
Upload:	andrew-ferguson
View:	214 times
Download:	0 times

Your Governance or Mine?

Documents