Presentation for Plenary

Session of the Annual

JCAE Symposium
8-9 January 2019

Investor Protection, Cross-listing

and Accounting Quality
Dr Khairul Anuar Kamarudin, Universiti Teknologi Mara
Dr Akmalia M. Ariff, Universiti Malaysia Terengganu
Prof Aziz Jaafar, Bangor University, UK
 Background
Benefits of Cross-listing
01
Outline Accounting Quality of Cross-Listed Firms

Literature Review
Investor Protection, Cross Listing and 02
Accounting Quality

Research Methods
Sample
Description of Data
03
Regression Models

Results
Descriptive Statistics
Regression Analysis
04
Additional tests
Discussion and Implications
 BACKGROUND
BENEFITS OF CROSS-LISTING

Access to fund
Greater access to capital for funding
growth opportunities (Hail & Leuz, 2009)

Governance aspect Reduce agency problem

Rent the host market’s superior More costly for controlling owners
corporate governance system and managers to extract private
(Licth, 2003) control benefits and to expropriate
outside investors (Stulz, 1999).

Better ‘information environment’

Greater attention by financial analysts and
monitoring by sophisticated U.S.
capital market participants
(Healy, Hutton and Palepu, 1999)
 BACKGROUND
ACCOUNTING QUALITY OF CROSS-LISTED FIRMS

‘Legal bonding’ is questionable as there are weak enforcement

mechanisms for foreign firms (Gande and Miller, 2012)

Diffusion between corporate governance

of the firm’s home and host environments
Accounting
Evidence of accounting scandals Quality of
involving cross-listed firms in the US Cross-Listed
Firms
Listing effects are exchange-specific
(Eng, Nabar, & Mian, 2008)

Influence of country-level features (Isidro, Nanda

& Wyoscki, 2016)
 MOTIVATIONS

• Shed light on how the relationship between cross-listing and accounting

quality varies across jurisdictions.

• Extend the work by Lang et al. (2003) by considering the effect of

investor protection.

• Add to the literature on bonding and signalling theory, suggesting the

need to explore beyond ‘legal bonding’ in dealing with cross-listed firms.
 LITERATURE REVIEW
INVESTOR PROTECTION, CROSS-LISTING AND ACCOUNTING QUALITY

Bonding Hypothesis Better information Home vs Host

environment Environment H1 Cross-listing in the US
Cross-listed firms
commit themselves to Firms cross-listed in the Cross-listed firms have stock market is positively
the stricter enforcement US have better
accounting quality
more tendency to associated to accounting
and litigation manage earnings,
environment of the host relative to foreign firms relative to the US firms
quality
market (Eng, Nabar, & that are not (Lang, (Lang et al., 2006).
Mian, 2008 ) Raedy & Yetman, 2003)
H2 The positive association
Managerial Incentive Regulatory Aspect between cross-listing in the
Institutional Theory
Managers’ incentives Rigorous regulation of USstock market and
Home countries factors
determine the quality of
for opportunistic the host countries accounting
behaviour decreases could have a different
financial reporting
with the level of regulatory burden for
quality is greater for firms from
(DeFond, Hung, & strong investor protection
investor protection firms in different
Trezevant, 2007)
(Leuz, Nanda, & institutional countries than weak investor
Wysocki, 2003) environments
protection countries
 RESEARCH METHODS
SAMPLE SELECTION

Sources of Data Final Sample

• Osiris-Bureau van Dijk
• Djankov, La Porta, Lopez- 32 countries
de-Silanes, & Shleifer 42,808 observations
(2008)
• La Porta, Lopez-de-Silanes, Missing Data & Outliers
Shleifer, & Vishny (1998) Removed missing data
Winsorized observations top and bottom 1%
Countries Excluded
1) with less than 100 observations
2) no US-listed firms

Industries Excluded
All financial institutions (SIC code 6000 - 6999) and
utilities companies (SIC code 4900 - 4999).
Sample Period
Year 2007–2016 .
 RESEARCH METHODS
DESCRIPTION OF DATA

Variable Measurements

• Timely loss recognition

- Asymmetric timeliness of earnings (Basu,1997)
- Accruals-based test of loss recognition (Ball & Shivakumar, 2005)

Accounting • Earnings management

Quality - Absolute & unadjusted value of residuals (ModifiedJones (1991)

• Value relevance
- The price model (Barth, Beaver & Landsman, 1998)
- The earnings model (Easton & Haris, 1991)
 RESEARCH METHODS
DESCRIPTION OF DATA

Variables Measurements
• Common vs Civil Law:
• Based on La Porta, Lopez-de-Silanes, Shleifer and Vishny (1998)

• Anti-director rights index :

Investor • Djankov, La Porta, Lopez-de-Silanes and Shleifer (2008)
Protection
Use a dummy variable HIADRI where value 1 is assigned for high investor
protection countries if the index is four or higher, and value zero for weak
investor protection if the anti-director rights index is below four.

• A dummy variable that takes value 1 if the firm cross-list to US market,

Cross-Listed
otherwise 0.
 REGRESSION MODELS
Timely loss recognition

EYit = β0 + β1RETit + β2RDit + β3RETit*RDit + β4CROSSit (1)

+ β5RET*CROSSit + β6RD*CROSSit + β7RET*RD*CROSSit
+ ψ1-nCountry_Effectsit + θ1-nYear_Effects + εit

ACCit = β0 + β1CFOit + β2DCFOit + β3CFOit*DCFOit + β4CROSSit (2)

+ β5CFO*CROSSit + β6DCFO*CROSSit + β7CFO*DCFO*CROSSit

where
+ ψ Country_levels + θ Year_effects + ε
1-n it 1-n it
EYit is the earnings yield, measured by earnings per share deflated by beginning of the fiscal year’s price per share ;
RETit is the annual return during the fiscal year;
RDit is a dummy variable equals to one (1) if RETit is negative, and zero (0) otherwise;
CROSS is a dummy variable equals to one (1) if a firm is cross-listed in the US stock market, and zero (0) otherwise;
ACC is measured as ∆Inventory + ∆Debtors + ∆Other current assets - ∆Creditors - ∆Other current liabilities - Depreciation deflated
by total assets at the beginning of the year;
CFO is operating cash flows over total assets at the beginning of the year; and
DCFO is a dummy variable that equals to one (1) if CFO < 0, and zero (0) otherwise.
 REGRESSION MODELS
Earnings Management

ABSDACC= α0 + β1CROSSit + β2AGEit + β3BIG4it + β4SIZEit + β5FMGROWit (3)

+ β6INDGROWit + β7LEVit+ β8CFOit+ β9CAPINTit+ β10ININTit
+ ψ1-nCountry_Effectst + θ1-nYear_Effectst + εit

DACC= α0 + β1CROSSit + β2AGEit + β3BIG4it + β4SIZEit + β5FMGROWit (4)

+ β6INDGROWit + β7LEVit+ β8CFOit+ β9CAPINTit+ β10ININTit
+ ψ1-nCountry_Effectst + θ1-nYear_Effectst + εit
where
ABSDACC is the absolute value of residual generated from the modified Jones (1991);
DACC is the unadjusted value of residual generated from the modified Jones (1991);
SIZE is the log of total assets;
LEV is total assets divided by total assets;
AGE is the log of number of years since the year of incorporation;
FMGROW is firm-specific growth, measured changes in sales compared to the previous year;
INDGROW is the mean of FMGROW calculated based on the Fama and French’s (1997) 48 industry groups;
CAPINT is the ratio of net book value of property, plant and equipment to total assets;
ININT is the ratio of research and development expenditure to total sales.
 REGRESSION MODELS
Value Relevance

PRICEit = α0 + β1BVit + β2E + β3CROSSit + β4BV*CROSSit + β5E*CROSSit (5)

+ ψ1-nCountry_Effectsit + θ1-nYear_effects + εit

RETit = α0 + β1Eit + β2CROSSit + β3E*CROSSit + ψ1-nCountry_Effectsit (6)

+ θ1-nYear_Effects + εit

where
PRICE is the closing stock price of firm; BV is the book value of equity per share;
E is earnings per share;
RET is the annual stock return;
RESULTS
DESCRIPTIVE STATISTICS

Cross-listed firms shows higher :

• stock prices (PRICE),
• stock returns (RET),
• book value of equity (BV),
• firm size (SIZE),
• earnings per share (E),
• earnings yield (EY),
• firm growth (FMGROW),
• average industry growth (INDGROW),
• operating cash flow (CFO)
• debt to assets ratio (DA),
• absolute value for discretionary
accruals (ABSDAC)

BUT lower
• log of age (AGE),
• value of unadjusted discretionary
accruals (DAC)

AND mostly audited by BIG4 audit firms.

 RESULTS
REGRESSION ANALYSIS : Asymmetric timeliness of earnings model

Models (1), (4), (5) : RET*RD*CROSS

20% Incremental timely loss recognition
incross-listed firms compared to
non-cross-listed firms in high
30% investorprotection environment and
the full
sample

50% Models (6) , (7) : RET*RD*HIADRI

and RET*RD*COMLAW
Incremental timely loss recognition
forcross-listed firms in high investor
60% protection countries than in low
investor protection countries
 RESULTS
REGRESSION ANALYSIS : Accruals-based test of loss recognition

Models (1), (4), (5) : CFO*DCFO*CROSS

20% Increase in the timely reporting of losses
in cross-listed firms compared to
non-cross-listed firms in high investor
30% protection environment and the full
sample

50% Models (6), (7) : CFO*DCFO*HIADRI

and CFO*DCFO*COMLAW
Existence of a more timely reporting of
losses in cross-listed firms from high
investor protection countries compared
60% to low investor protection countries
 RESULTS
REGRESSION ANALYSIS : Absolute Discretionary Accruals (ABSDACC)

Models (1) to (5) : CROSS

20% Cross-listed firms have greater
tendency to manage earnings as
compared to their domestic
30% counterparts in both low and
high investors protection countries

50% Models (6), (7) : HIADRI and

COMLAW
Cross-listed firms in high investor
protection countries have lower
tendency to manage earnings
60% compared to cross-listed firms in
low investor protection countries
 RESULTS
REGRESSION ANALYSIS : Unadjusted Discretionary Accruals (DACC)

Models (1), (3), (4), (5) : CROSS

Cross-listed firms in full sample, high
20% anti-director rights sample and both
types of legal system pursue earnings
decreasing strategy

30%
Model (6) : HIADRI
No evidence in difference in earnings
management direction between high and
50% low anti-directors rights countries

Model (7) : COMLAW

Cross-listed firms in common law
60% countries have stronger tendency to
manage earnings upwards
compared to cross-listed firms in civil
law countries
 RESULTS
REGRESSION ANALYSIS : Price Model

Models (1) to (5) : BV*CROSS and

E*CROSS
20%
Higher value relevance of accounting
numbers in cross-listed firms compared
to non-cross-listed firms.
30%

Models (6), (7) : E*HIADRI and

50% E*COMLAW

Value relevance of earnings in

cross-listed firm is higher in low investor
protection countries than in high investor
60% protection countries.
 RESULTS
REGRESSION ANALYSIS : RETURN MODEL

Model (1), (2), (3), (5) : E*CROSS

20% Significant increases in value relevance
of earnings in cross-listed firms,
especially in low anti-directors rights and
weak legal system samples
30%

Models (6) and (7) : E*HIADRI and E*COMLAW

50%
Higher value relevance of earnings
in cross-listed firms in low investor
protection countries than high
investor protection countries.

60%
 ADDITIONAL ANALYSES
ROBUSTNESS

YEAR AUDITOR’S QUALITY LARGE SAMPLE

ALTERNATIVE
2008-2011 Big 4 auditors . Exclude UK and
MEASURES
global crisis / Japan
• ASDI (Djankov et al,
Rule 12h-6
2008)
• IPWEF (World
Economic Forum)

Endogeneity
ROBUSTNESS
Two-stage estimation
TESTS procedure proposed by
Heckman (1979) that is by
using the inverse Mills ratio
to correct for the selection
bias
 Endogeneity
First Stage Second Stage
Timely Loss Recognition Earnings Management Value Relevance
Probit
Asymmetric timeliness of earnings (Absolute discretionary accruals) (Price model)
Model (1) Model (2) Model (3) Model (4)
Intercept 11.440*** Intercept 0.241*** Intercept 0.100*** Intercept -2.846
(31.601) (3.608) (10.088) (-0.403)
GDP -1.399*** RET 0.051*** HIADRI -0.022** BV 1.118***
(-42.243) (3.255) (-2.322) (33.662)
DISC 7.477*** RD 0.027 AGE -0.002*** E 12.351*
(72.182) (1.471) (-4.709) (1.693)
IFRS -0.807*** RET*RD 0.181*** BIG4 -0.002 HIADRI -0.023
(-26.419) (3.597) (-1.458) (-0.497)
OWN -1.158*** HIADRI -0.145** SIZE -0.004*** BV* HIADRI 7.035***
(-11.072) (-2.111) (-14.600) (32.030)
RET*HIADRI -0.030* FMGROW 0.000 E* HIADRI -0.949***
(-1.738) (0.124) (-2.998)
RD*HIADRI 0.002 INDGROW 0.001*** IMR -0.641
(0.100) (2.839) (-0.261)
RET*RD*HIADRI 0.143*** LEV 0.002***
(2.628) (5.196)
IMR 0.021 CFO -0.001
(0.896) (-0.238)
CAPINT 0.016***
(8.316)
ININT 0.000
(0.649)
IMR 0.002
(0.468)
 Agenda
RESULTS Style
TIMELY RECOGNITION OF LOSSESS
SUMMARY
The level of investor protection has influenced the cross-listing effects on
earnings conservatism.

EARNINGS MANAGEMENT
Cross-listed firms are associated with earnings decreasing strategy, except for
the low anti-directors rights sample.

VALUE RELEVANCE
The positive impact of cross-listing on the value relevance of book value of equity
and earnings. This is more pronounced in low investor protection countries

Cross-listing is associated with enhanced accounting quality but the

association is influenced by the extent of home country institutional
factors.
IMPLICATIONS Agenda
Literature
• Bonding and signalling theory
- The need to explore beyond ‘legal
bonding’ in dealing with cross-listed
firms
• Cross-listing literature
- Focus beyond the host-country
environment
• Institutional theory
- Cross-listed firms face
incentives from the various
institutional factors of the home
country
Style Policy
• Standard setters
- Convergence in corporate
governance and the future of
capital market integration
• US policy maker
- Long-term strategy for consistent
accounting quality of the foreign
firms listed in the US.
• Home-country’s policy maker
-Stronger institutional features of
the home country
 Thank you
Questions and suggestions are highly appreciated.