ifrs on australian firms

7/26/2019 IFRS on Australian Firms

1/31

The Effects of International Financial Reporting

Standards on the Accounts and Accounting

Quality of Australian Firms:

A Retrospective Study

John G o o d w i n * ,

Kamran

A h m e d b *

and

Richard H e a n e y c *

TheHong

Kong

Polytechnic University

bLaTrobe University

'RMIT University

Received September 2007; Accepted June 2008

Abstract

We exam ine the effect of Australian equivalents to International Financial Reporting Standards

(IFRS)

on the accounts and accounting quality of

1,065

listed firms, relying on retrospective

reconciliations between Australian Generally Accepted Accounting Principles (AGAAP) and

IFRS. We find that IFRS increases total liabilities, decreases equity and more firms have earnings

decreases than increases. IFRS earnings and equity are not more value relevant than AGAAP

earnings and equity and while adjustm ents for changes in accounting fo r provisions and intangibles

other than goodwill are value relevant, they weaken associations with market value. Goodwill

adjustments improve associations with market value. We also find that the reconciliation note for

the earnings adjustments contained no new information.

JE L C lassi3cations:

M40

41

Keywords: LFRS, accounting quality

*We hank participants at research w orkshops at The H ong K ong Polytec hnic University,

RMIT

University,

The University

of

Technology Sydney, the AAA

2007

annual meeting, the AFAANZ 2007 annual meeting and

the JCAE/AJPT

2008

annual symposium for their valuable input. We also thank Eli Bartov, Kim Sawyer, Katherine

Schipper, the editors and the anonymou s reviewer

for

their valuable input. Any errors remain the authors'.


2/31

90

John Good win, Kamran Ahmeda nd Richard Heaney

Journal of Con temp orary Accounting Economics

Vol4 ,

No 2 (December

2008

89-119

1.

Introduction

In 2002 the Financial Reporting Council (FRC) claimed that the implementation

of Australian equivalents to International Financial Reporting Standards (IFRS) would

enhance the overall quality of financial reporting in Australia FRC,002). This view

was supported by the Australian Accounting Standards Board (AASB) in suggesting that

complying with IFRS would not impose significant burdens and costs on entities when

compared with the benefits of more relevant and reliable information for users of financial

reports (Fenton-Jones, 2003,53). There was general agreement among commentators, the

firms, analysts and the wider community that the introduction of

IFRS

from 2005 would

materially affect Australian firms financial performance and accounts quality (Buffini,

2005; Clarke and Dean, 2005, Deegan, 2005,32-35). A study based on interviews of 60

senior financial executives from Australias top 200 firms reported that the introduction

of IFRS would have a significant impact on financial position and earnings. The majority

of executives expected earnings to be negatively affected while less than half expected a

positive effect on the financial position of their firms (Jones and Higgins, 2006). However, no

substantial empirical study has been undertaken to assess the claim that financial statements

prepared under IFRS will enhance the quality of financial reporting in Australia or to

examine the effects of IFRS on financial performance. In this paper we use a representative

sample of listed firms in attempting to fill this gap in the literature.

Our study contributes to the current debate on whether IFRS based accounting numbers

are of a different quality to those produced under domestic GAAP in several important

ways. First, we examine all listed Australian firms that have available data. Second, the

exemptions from applying IFRS to restated earnings and equity are limited to a small

number of standards and all firms are required to restate and provide reconciliations from

Australian Generally Accepted Accounting Principles (AGAAP) to IFRS upon first-time

adoption of IFRS (AASB 1 First-Time Adop tion ofAustralian Equivalents to International

Financial Standards,

para 39): Prior studies use datasets that may not be representative

of

the full effects of

IFRS

due to small sample size (e.g., Hung and Subramanyam, 2007)

or they may use firms that voluntarily adopt IFRS (e.g., Barth et al., 2005; Bartov et al.,

2005), meaning control for self-selection bias is needed. Self-selection bias is not an issue

with our dataset because early adoption is not permitted (AASB

1).

Third, our large sample

size permits an empirical examination

of

the reconciliation adjustments to IFRS, which

has not been reported in the literature. Finally, examining the switch to IFRS for Australia

is useful, as similar studies have focused on code law countries such as Germany. In

Germany, accounting numbers are more conservative and have different value relevance

than accounting numbers produced under common law based countries like Australia (Ali

and Hwang, 2000; Ball et al., 2000).

IFRS is applicable for reporting periods beginning after 31 December 2004 and firms

are required to restate comparatives and provide reconciliations o IFRS in their notes to the

Supporting this contention are the shareholder briefings

held by

som e firms (e.g., Telstra and Alinta)

to

explain the financial impact

of IFRS

on their accounts.

There are some standards exempt from retrospective application, namely the two financial instruments

standards (AASB 132 and AASB 139),AA SB 3

Business

Combinationsand the three insurance standards

(AASB

4 nsurance Con tracts,AASB 1023General Insurance Contractsand AASB 1038

Life

Insurance Contracts) .


3/31

John Good win , Kamran Ahmedand Richard Heaney

91

Journal of Contemporary Accounting Economics

VoE4,

No 2 (D ecember 2008)

89-119

accounts in the first year of adoption. This requirementpermits a research design that directly

compares accounting numbers and their properties prepared under AGAAP with those

under IFRS for the same set of firm-years, as firms provide earnings and equity amounts

measured under two different sets of accounting standards for the same periods.

Our investigation comprises two main parts. First, we document the effect of IFRS on

key accounting numbers and ratios. Using a sample of 1,065listed firms, we find that the

mean (median) of total liabilities has increased and the mean (median) of total equity has

fallen. Total assets and earnings are higher under IFRS but the changes are not significant

apart from the increase in the half-year earnings median. IFRS increases the leverage ratio.

Second, we examine the relative value relevance of

IFRS

earnings and equity and the

incremental value relevance of IFRS over AGAAP. Using models with market prices and

returns as dependent variables, we carry out our tests on annual earnings (net income) and

equity measured at the changeover date to IFRS. We find no evidence that

IFRS

earnings

and equity are more value relevant than AGAAP. We find weak evidence that the aggregate

changes for earnings and equity are incrementally value relevant to AGAAP. About half of

the reconciling adjustments to AGAAP are value relevant but are generally not timely. The

intangibles and provisions adjustments in particular weaken the association with market

value. Goodwill adjustments improve the association with market value.

Because our data is from retrospective reconciliations, the tests are only of the ability

of accounting numbers to capture information used by the market. Therefore a coefficients

significance provides a lower bound on its true significance. We are aware of two other

similar papers, namely Rees and Elgers (1997) and Hung and Subramanyam (2007)

that examine a retrospective dataset. In our final test we examine whether the earnings

adjustments are associated with market values measured over the interval covering

the

release of the IFRS reconciliation note. This test permits inferences about the usefulness

to investors of the reconciliation adjustments, as observing a significant coefficient in

this regression could result from investors using that information. We find that all of the

information in the earnings adjustments was impounded into prices before the release of

the accounts, suggesting that the market was able to obtain the information from sources

other than the reconciliation note.

The remainder of this paper is organised as follows. The literature review is covered

in the next section. Section

3

describes the institutional background and data collection

process. Section 4 describes the accounting differences between AGAAF and

IFRS

for the

major reconciliation adjustments and provides an examination of the effects of IFRS on

the accounts and on financial statement elements and ratios. Section5 covers methodology

and results for relative value relevance tests and incremental value relevanceis covered in

section

6.

Section 7 concludes the paper.

2.

Literature Review

Barth et al. (2005) use data from 24 countries over a 15-year period to 2004 and find that

the transition to IFRS esults in improved accounting quality using a variety of measures.

Specifically, they find that IFRS results in more timely recognition of losses and higher

R2s in regressions of market value on earnings and book value. A study using European

data by Armstrong et al. (2007) report that the stock market reacts positively to the early


4/31

92

John G oodwin, Kamran Ahmed and Richard Heaney

Journal

of

Contemporary Accounting Econom ics

V o l 4 ,N o 2

(December 2008)

89-119

adoption of IFRS, which suggests that European equity investors perceive net benefits due

to convergence of accounting standards and improved information quality following IFRS

adoption.While these results may be more representative of the overall effects of IFRS,they

are difficult to interpret for a study examining one country because each set of domestic

GAAP is likely different in a number of respects. For example, the timeliness characteristic

of earnings differs according to the institutional environment between countries (Ball et

al., 2000): The most closely related studies to the present study are those that examine

samples from one country.

An early single-country study by Kinnunen et al. (2000) exploits a unique market

setting in Finland, where foreign investors are restricted in their trading of certain shares.

This permits the authors to examine the relative value relevance of Finnish GAAP and

voluntarily adopted IFRS between two investor groups. They find IFRS improves the

information content for foreign investors but not for domestic investors. Another Finnish

study by Niskanen et al. 2000), examines components of reconciliations to IFRS for

18 Finnish firms voluntarily using IFRS over the period 1984 to 1992. They report the

aggregate earnings difference is value irrelevant for explaining returns but that untaxed

reserves adjustments and consolidation differences are value relevant. Since these papers

examine voluntary adopters the results may be affected by self-selection bias and they use

a dataset of accounting rules that is dated

More recently, Christensen et al. (2007) have examined the economic consequences for

UK firms of the European Unions decision to impose mandatory IFRS. They show that

there are cross-sectionalvariations in both short-run market reactions and long-run changes

in cost of equity associated with the decision, suggesting that mandatory IFRS adoption

does not benefit all firms in a uniform way but results in winners and losers. Using a price

levels regression, Hu (2003) reports that Chinese GAAP is more value relevant than IFRS

using a sample of 252 firm-years. This finding is supported by Eccher and Healey (2003),

who investigated a sample of 83 Chinese firms that were required to provide two sets of

accounts using Chinese GAAP and

IFRS,

and found that earnings under Chinese GAAP

were more closely associated with returns than earnings under IFRS.

Hung and Subramanyam (2007) use a sample of 80 German firms which voluntarily

adopted IFRS over the period 1998-2002 and provided accounts under German GAAP and

IFRS for the same period. Using price levels models, they find that total assets and book

value of equity, as well as variability of book value and net income, are higher under IFRS

than under German Accounting Rules (HGB). They also find that book value of equity

and net income under IFRS are no more value relevant than the amounts under HGB.

Further, they report that earnings and equity under IFRS are incrementally value relevant

With respect to a countrys institutional environm ent, Ball et al. (2000) report that du e to different levels

of conserv atism, earnings of firms in common law based countries are more timely in impounding ec onom ic

information than earnings of firms in co de law countries. They al so find that earnings co efficients are larger for

code law than for common law countries.

Other studie s on the voluntary adoption

of

IFRS investigate the distinguishing eco nom ic characteristics of

firms that switch to IFRS.El-G azza r et al. (1999) report that firms voluntarily adopting IFRS are those seeking

to access foreign capital, improve customer recognition or reduce political costs. Lang et al. (2003) note that

firms electing to adopt IFRS early are more likely to be those firms with fewer reconciling item s.


5/31

John Goodw in, Kamran Ahmedand Richard Heaney

93

Journal of Contem porary Accounting Econom ics Vol4 ,No

2

(December 2008 89-119

to German GAAP. Both coefficients are highly significant but the earnings coefficient sign

is negative which they suggest is consistent with more measurement errors in the

IFRS

earnings than in the German earnings. Their relative value relevance results contrast with

Bartov et al. (2005) who also compare German HGB with

IFRS

(and

US

GAAP). They

find that

IFRS s more value relevant than German HGB in its ability to explain returns (as

opposed to prices used by Hung and Subramanyam,2007). They also find little difference

in the value relevance of US GAAP earnings and IFRS earnings after self-selection bias

is controlled for?

Bartov et al. (2005) examine firms in the cross section rather than the same set of firms

as do Hung and Subramanyam (2007). The present study also examines the same firms with

the added advantage that firms must adopt IFRS rather than voluntarily. Although Hung

and Subramanyam (2007) control for self-selection bias, the possibility remains that the

bias in the

IFRS

accounts due to the effects of early voluntary adoptions may explain the

conflicting results, as Barth et al. (2005) have noted. Different models might be another

reason, which motives us to estimate both a price levels model and a returns model.

In sum, the literature gives conflicting evidence on accounting quality. Our study adds

to the existing literature on the effects of adopting

IFRS

on earnings and equity quality by

using recent

data

rom companies that are required at initial adoption to provide earnings and

equity numbers under both AGAAP and IFRS,and by using price and returns models.

3.

Institutional Background and Data

Australian firms are required to use IFRS for reporting periods beginning on

or

after

1

January 2005. As part of the transition to IFRS,AASB

1

requires retrospective

reconciliations to

IFRS

earnings and equity to be provided in the first half yearly accounts

using IFRS and the first annual accounts using IFRS. or instance, in its annual accounts

ended 30 June 2006, the firm is required to provide a reconciliation of its annual AGAAP

earnings to

IFRS

earnings for the year ended 30 June 2005 and

of

its AGAAP equity to

IFRS

equity at 30 June 2005

6

These data permit an examination of the major differences

between AGAAP and IFRS earnings and equity both in aggregate and for specific

reconciliation adjustments.

Our final sample of 1,065 firms is determined in a series of steps. First, all firms listed

on the Australian Stock Exchange are selected from Aspect Huntley Pty Ltds FinAnaZysis

database as at January 2006 which gives a total of 1,714firms. From these, 180are deleted

because they were listed in 2005 or 2006; 72 because they used non-Australian GAAP

or foreign currency; 73 because they were suspended from trading and their accounts

are

not available; and two because they were delisted prior to account lodgement. This gives

a potential sample of 1,387 firms. We then examine the notes to the annual accounts and

Leuz 2003)

also examines the value relevance of German accounting numbers and reports that neither the

trading spread nor the trading volumes is significantlydifferent for the companies that choose IFRS or US GAAP

voluntarily.

Another requirement is to provide a reconciliation to half-year IFRS earnings (net income) to December

2004 in the half-year accounts ended

3

1 December2005. These dates are different for afirmwith a year-end that

differs from June

30

(see paragraphs

39

and

45

of

AASB

1 for other requirements).


6/31

94


Journal

of

Contem porary Accounting Econom ics

Vol4 ,

No 2 (December 2008)

89-119

record details of any reconciliations for annual net income and equity at the last date that

AGAAP

is used.

We delete 280 firms which were unaffected by IFRS and 42 firms with missing data

or identified as outliers in initial regression tests. In the final sample of

1,065

firms

there

are

1,020

firms with earnings changes and 844 firms with equity changes. Table

1

shows

the sample description.

Table

1

Sample Description

Numb er of firms listed on ASX at January 200 6 1,714

Less:

Firms listed in 2005

or

2006

(180)

Accounts not available (73)

Firms delisted in 200 6

2)

F m s

using foreign currency / foreign GAAP

(72)

Potential sample 1,387

Financial 18% 253

Non financial 47% 649

1,387

Mining 35%

85

Less:

Unaffected by IFR S

(280)

Firms with missing data

Final sample 1,065

Financial 18% 193

Non financial 50 530

1,065

Mining 32%

Firms with earnings chang es

1,020

Firms with equity change s 844

4.

Major Accounting Differences and Their Effects on the Accounts

In this section we describe the main accounting differences between AGAAP and IFRS

for major adjustments, discuss the main reasons for each adjustment, and present evidence

of

their effects on the accounts.

4.1

Ma jor accounting differen cesand adjustments

Table 2 shows frequencies and percentages of unsigned

AGAAP

earnings and equity of

the nine most common reconciling adjustments for earnings, and for a catch-all component

Of the

280

unaffected firms,

15

firms changed comparatives only for prior period

errors

and accounting

policy changes unrelated to IFRS .AASB 108Accounting Policies, Cha nges in Accounting Estimates and Errors

requires retrospective adjustment for these types of change (para 19and 42).


7/31

John

Goodwin,

Kamran Ahmedan d R ichard Heaney

95

Journal

of

Contemporary Accounting Economics

Vol4 ,

No 2

(December

2008)

89-119

Table2

~

Frequencies and Percentagesof Unsigned Earnings and Equity for Adjustmentsof Earnings and Equity

Earnings Equity

N

Mean Median

N

Mean Median

Share-based payment

Incom e tax

Goodwill

Intangibles

Provisions

Investments

Impairment

FX

translation

Leases

Other

Total

594

456

381

183

165

131

98

98

90

353

1 020

-34

40

9

-3

75

-54

-4

-4

-16

-1

-3

0

9

1

-1

6

0

0

0

- 1

0

83

41

315

194

174

94

120

84

84

361

844

-4

0

3

-11

-1

5

-12

-1

-1

-9

-6

-1

0

1

-1

0

0

-4

0

0

-1

-1

Earnings is annual net income . Equity is net assets. Mean (median % )is the mean (median) of the reconciling

adjustment amount

for

earnings or equity divided by the absolute value of AGAA P earnings or AGAA P equity

for each firm. ther is a catch-all comp onent.

-

other? A summary of accounting policy differences between AGAAP and

IFRS

is

shown in Appendix A.

Share-basedPayment:

Under

IFRS ,

AASB 2

Share-based Payment

requires expensing

of the fair value of equity granted. The prevalence and effect of this adjustment is due

to the popularity of equity schemes and the almost complete lack of equity payment

expensing prior to

IFRS.

The recognition of the expense for some small

firms

is highly

material (untabulated), thus the mean of -34 percent is much larger than the median of -3

percent. Because the accounting is equity neutral, the number of adjustments to equity is

much smaller than for earnings. The negative adjustments to equity are mainly caused by

derecognition of loans to employees to acquire shares and derecognition of the entitys own

shares held in trust for employees (treasury shares) under AASB

132

Financial Instruments:

Disclosure and Presentation.

We classify these types

of

treasury share adjustments here

because it presents a clearer picture on the effects of employee share scheme accounting.

Other adjustments for treasury shares are included in financial instruments below.

Income

Tax:

The balance sheet method of income tax accounting is used under

AASB

112 Income Taxes

compared with the income statement method under AGAAP.

The principal difference is that the tax effects of items that bypass the income statement

are not recognised under AGAAP. Consequently, common adjustments relate to asset

revaluations, which give rise to deferred tax liabilities (DTLs), and share issue costs in

share capital, which give rise to deferred tax assets (DTAs). Raising DTAs for the

tax

effects of share-based payments expenses is common since these expenses were usually

unrecognised under AGAAP, hence the positive mean. DTLs recognised for intangibles

* T h e

irms

disclosures are used to guide classifying the adjustments.


8/31

96

John Goodwin,

Kamran

Ahmedand Richard Heaney

Journal of Contemporary Accounting Economics

V o l 4 ,N o 2

(December 2008)

89-119

that have indefinite lives are also quite common.

Goodwill:

Amortisation over a period not exceeding 20 years under AGAAP is

replaced by an annual impairment test under AASB 3 Business Combinations.The result

is that firms have an increase in earnings and in equity on transition to IFRS. Hence the

means and medians are positive for earnings and equity in Table 2. Several firms wrote

off the balance of goodwill at transition due to impairment and these are classified under

impairment below. As with share-based payments, small firms are the cause of the large

mean of 40 percent relative to the median.

Intangibles: Under AGAAP, acquired and internally generated intangibles can be

capitalised and revalued provided that the value is reliably estimable9. Under AASB

138 Intangible Assets only acquired assets and development costs can be capitalised and

revaluation is only available for assets with an active market, with the result that most

intangibles are carried at cost. Assets are amortised under both AGAAP and IFRS although

amortisation can be avoided under

IFRS

if the asset has an indefinite useful life.

Except for development costs, it is clearly more difficult to capitalise intangibles and

measure them at fair value under IFRS than under AGAAP. Under IFRS development

costs can be capitalised provided the benefits are probable

lo

Under AGAAP, research and

development (R&D) costs can be capitalised provided the benefits are beyond reasonable

doubt, which represents a higher hurdle than probable. However, there are specific

criteria to be met for capitalisation under IFRS, meaning the change in reporting bias is

ambiguous for these assets.

As

a

result of applying AASB 138 retrospectively, some firms have derecognised

their R&D asset and reversed its associated amortisation; and other firms have capitalised

development costs, which they had immediately expensed under AGAAP. In the first case,

at transition to IFRS the firm has a negative equity effect and typically a positive earnings

effect, and in the second case those firms have positive earnings and equity effects. Most

adjustmentske of the first kind, thus Table2 shows positive median adjustmentstoearnings

and negative median adjustments to equity, most of which relate to R&D. Derecognition

of other intangible assets occurs because those assets are not measured by reference to

an

active market. Examples include brand names, trademarks and publishing rights and titles,

which can be very large, hence the mean of -11 percent.

Provisions:

This adjustment includes initial recognition of costs restoration provisions

for leasehold improvements and assets under operating leases under AASB 137Provisions,

Contingent Liabilities and Contingent Assets. Previously, these costs were recognised on a

cash basis except for mining firms, but under AASB 116Property, Plant and Equipment,

the cost of an asset includes the cost of restoration. The re-measurement of AGAAP

provisions for mining firms on a discounted basis gives positive equity adjustments for

these firms, although the fair valuing of provisions under

IFRS

generally leads to negative

adjustments. Other positive adjustments are reversals of general provisions disallowed

under AASB 137, such as doubtful debts and restructuring provisions typically recognised

in a business combination, and reversals of provisions for dividends.

Recognition

of

internally generated goodwill was prohibited under

AGAAP.

Mining firms measured the provision on an undiscounted hasis however.

o Probable is generally accepted to be more than 50 percent likely.


9/31

John Goodwin, Kamran Ahmedand Richard Heaney

Journal

of


Vol4 , No 2

{December2008) 89-119

91

Investments: Revaluations of property are recognised in the income statement under

AASB 140 Investment Property, but under AGAAP such revaluations were recognised

directly in equity. This is the underlying reason for the large difference between the

percentage adjustment to earnings

( 7 5 )

and equity

( 5 ) .

Also included in this adjustment

are re-measurement of investments in associates due to the collective effect of

IFRS,

and

equity investments classified as available for sale. Most adjustments are positive, but the

few negative adjustments are due to initial depreciation on investments reclassified to

property, plant and equipment, because those investments do not meet the definition of

investment under AASB 140.

Impairment: Under AASB 136 Impairment

of

Assets, firms are required to assess

non-current assets to determine whether the carrying amount exceeds the recoverable

amount, measured at present value. Under AGAAP, firms could choose to measure the

recoverable amount using undiscounted or discounted cashflows. More firms and firms

with large write-downs recognised impairments against opening equity at transition, as the

median of

-4

ercent for equity versus a median of

0

percent for earnings suggest.

FX

Translation:Under AGAAP, firms used either the current rate or temporal methods

to translate foreign operations accounts. Under AASB 121 The Effects

of

Changes in

Foreign Exchange Rates , firms are required to determine their functional and presentational

currencies and translate into Australian dollars at the rate current at balance date. The

result is that foreign exchange gains and losses on translation are now recognised directly

in equity until the investment is sold when the net of those gains and losses is recognised

in earnings. Non-monetary items are now translated at the current rate.

Leases: The main change under

IFRS

is the requirement to measure on a straight line

basis the expense for operating leases that contain payment escalation clauses. Previously

under AGAAP, lessees and lessors measured the expense on a cash basis. Lessees generally

have negative adjustments o earnings and equity indicating underexpensing in prior years.

Both lessees and lessors are classified here,

so

some adjustments to earnings are positive,

but most are negative.

Other: Other is a plug component necessary to complete the reconciliation from

AGAAP to

IFRS.

Included here are adjustments for financial instruments, prior period

errors and financial effects of changed accounting policies unrelated to

IFRS.

The bottom row of Table

2

shows the overall mean (median) decrease to earnings

is 7 0)percent and to equity it

is

6(1) percent, consistent with a more conservative bias

under

IFRS.

4.2

Effect of changes on jn an cia l statement elements and ratios

We

next examine the effect

of IFRS

on financial statement elements, retained profits,

cash flow and ratios. The top three rows of Table 3 shows that assets, liabilities and equity

differ under IFRS, but

only

the liabilities and equity differences are significant at the 10

percent level. The mean of total assets has risen yet the median is unchanged, due to some

very large asset increases for a small number of firms. These increases are mainly due

to first time consolidation of mortgage trusts by financial industry firms, which has also

increased liabilities with a zero or small net equity effect. By contrast, the effect is clearly

upward (downward) with liabilities (equity) due to recognition of new liabilities, such

as deferred tax liabilities for asset revaluations and restoration provisions; remeasuring


10/31

98

John

Go odw in, Kamran Ahmedand Richard Heaney

Journal of Contemporary Accounting Economics V o l4 ,No 2 (December 2008)

89-119

Table

3

Effect of IFRS on Financial Statement Numbers and R atios

Median

ean

N AGAAP IFRS t-test AGAAP IFRS Wilcoxon

Total assets

Total liabilities

Total equity

Retained profits

Earnings

Earnings - half year

Net cashflow

Operating cashflow

ROE

ROA

TL TA

Price earnings

Market to book

1 065

1 065

1 065

1 065

1 065

1 064

1 065

1 065

1 033

1 065

1 065

496

1 033

517.60

296.40

22 1.20

23.70

22.90

12.50

6.50

30.40

-0.24

0.38

21.70

2.76

-0.15

529.30

342.30

187

OO

13.60

24.40

1

3.20

6.10

3

1.40

-0.23

-0.16

0.41

21.56

3.23

0.49 30.30

30.30

0.02 7.80 8.20

0

oo

19.50 18.50

0.01

-3.90

-4.90

0.11 0.10

0.12

0.21 -0 o I

0

oo

0.27

0.25 0.25

0.21 -0.07

-0.07

0.73 0.01 0.02

0.19 0

oo

0.01

0.oo

0.31 0.33

0.97 10.95

10.51

0.12 1.71

1.84

0.34

0

oo

0

oo

0 oo

0.62

0.01

0.35

0.14

0.56

0.00

0

oo

0 oo

0.00

All balance sheet numbers are measured at the end of the last financial year reported under AGAA P. All

financial statement numbers are expressed in millions. RO E is earn ings divided by equity at year-end. RO A

is annual earnings divided by total assets a t financial year-end. TL is total liabilities. TA is total assets. Price

earnings ratio is calculated only for positive earnings for both AG AA P and IFRS. OE nd market to book ratio

are calculated on smaller samples because some firms (unincorporated trusts) report zero equity under IFRS.

P-values are for two-tailed tests

existing liabilities more conservatively, such as defined benefit superannuation plans; and

restoration provisions and reclassifications, such as equity to debt for trusts.I2

Both yearly and half yearly earnings have increased under IFRS,but the differences are

insignificant except for the half-year medians. Despite these trends more firms experience

an earnings decrease than an increase. This occurs because the dollar amount of the change

to earnings is about four times larger when it is positive. The main reasons for this are

unrealised gains

on investments recognised in earnings under IFRS whereas those gains

were recognised directly in equity under AGAAP, and the reversal of goodwill amortisation.

The differences between operating and net cash flows are not significantly different.

With respect to the ratios the differences in means are insignificant with the exception

of leverage

(TL/TA).

The leverage median difference is also significant and indicates higher

accounting risk under

IFRS.

he ROA median is higher underIFRS, nd the price earnings

ratio, which is measured only for positive earnings, is lower. The median difference of the

market to book ratio is higher reflecting much lower equity values. The results shown in

Tables 2 and 3 indicate that

IFRS

presents a weaker balance sheet for the average firm.

l 2 This latter reason is not expected to impact fu ture reporting periods becau se trusts are presently amending

their deed s to circumvent the accounting requirement for trusts with limited lives to report ze ro equity.


11/31

John G oodwin, Karnran Ahm edand Richard Heaney

99

Journal

of


Vol4 ,

No 2 {December

2008)

89-119

5.

Relative Value Relevance

5.1

Methodology

Our second main objective is to examine the quality of IFRS-measured earnings and

equity. We adopt a value relevance perspective for accounting quality recognising that it is

but one of several perspectives (Schipper and Vincent,2003 .13Following similar studies

(e.g., Hung and Subramanyam,

2007),

a price levels model based on Ohlson

(1995),

and

a returns model (e.g., Bartov et

al.,

2005) are used. In this section we compare in the cross

section the value relevance of earnings and equity measured using AGAAP and IFRS for

firms that report a net change from IFRS. These tests determine the relative abilities of

AGAAP and

IFRS

earnings and equity to capture information used by the market.

The price levels modelI4 s:

where MV is the market value of the firms equity at the end of the last year that AGAAP

is used scaled by the number of shares at that time; E is the firms earnings for the last year

that AGAAP is used measured under AGAAP (denoted EA) or

IFRS

(denoted

EI)

scaled

by the number of shares at the end of that year; BVi, is the firms equity at the end of the

last year that AGAAP is used measured under AGAAP (denoted BVA) or

IFRS

(denoted

BVI) scaled by the number of shares at that time; and

E

is the error term.

We are also interested in examining the timeliness of the information contained in

AGAAP and IFRS-measured earnings. Our second model is:

where

R

is the annual raw return adjusted for capitalisation changes and dividends

measured to the end of the last year that AGAAP is used; E is the firms earnings for the

last year that AGAAP is used measured under AGAAP (denoted EA) or IFRS (denoted

EI)

scaled by the market value of equity at the start of that period; and E is the error term.

A

more common model includes earnings change as an additional independent variable

but data limitationsprevent us from using that model in the cross sectional comparison.We

do estimate that model in robustness tests in a longitudinal comparison (discussed below)

and inferences are unchanged.

Ball (2006)notes there is little settled theory or empirical evidence on which to build an assessment of

the advantages and disadvantages of

IFRS

despite the fact that almost

100

countries have adopted

them.

I

Som e studies (e.g., Easton, 1998;Easton et al., 2003) criticise scaling by the number of shares, arguing

that it may lead to spurious results due to the effects of scale. Barth and Clinch (2001) investigate the relative

efficiency of deflating methods in m itigating a scale effects model and find that

the

number

of

the firms outstanding

shares, as used in o ur study, is more efficient in m itigating sca le effects.


12/31

100

John G oodw in, Kumran Ahmedun d Richard H eaney

Journal of Con tempo rary Accounting

14

conomics V o l 4 ,

No 2

(December

2008

89-119

Because earnings releases contain information (Ball and Brown, 1968),value relevance

researchers often measure market value after the release of the accounts (e.g., three months

after fiscal year-end). However, in the present study this would bias the tests in favour of

AGAAP since the last AGAAP accounts are released before the measurement of market

value, and the IFRS reconciliation data are not released until the first IFRS accounts, which

is some months thereafter. Therefore, to place the two regressions on a more equal footing,

returns (and price) are measured at the end of the last year that AGAAP is used. Further,

as AGAAP half yearly earnings are released during the return interval used in Model ( 2 ) ,

which may introduce a bias favouring AGAAP, we also compare half yearly earnings using

model

( 2 )

and use 6-month returns. Results (untabulated) from this regression, discussed

in the robustness tests section below, leave inferences unaffected.

We assess accounting quality by comparing the explanatory power of the models

measured by the adjusted R2,using the test proposed in Vuong

(1

989). AGAAP regression

results are presented firstsoVuongs (1989) Z-statistic ispositive when the IFRS regression

is favoured over the AGAAP regression.

Table 4

Descriptive S tatistics for Regression V ariables

Panel A:

Price levels regression (Model

1 )

Code N Mean Median Std De v Min Max 70Pos

MV

MV

1,065

1.18

0.42

1.71

0.01

8.86

100

AGAAPequity BVA

1,065 0.73

0.23

1.14 -0.56

8.65 98

IFRS equity

BVI

1,065 0.66 0.20 1.04 -0.62 8.65 98

AGAAPeamings EA

1,065 0.05 0.00 0.19 -1.42 3.01 51

IFRS earnings EI

1,065 0.06 0.00 0.19 -1.02 3.01 51

Panel B:

Returns regression (M odel

2)

Returns R

922

0.22 0.05

0.92

-0.95 7.84 55

AGAAP earnings EA

922 -0.07 0.01

0.45

-8.82 1.68 52

IFRS earnings El

922 -0.06

0.02

0.37 -4.68

2.06 52

I n Panel A, all variables are scaled by number of shares outstanding at the end of the last year that AGAAP is

used.

In Panel B , returns are measured over the 12-month period ending on the last day that AGAAP is used and

all independent variables are scaled by market value

of

equity at the start of the last year that AGAAP is used.

Detination of variables are described as in the text.

5.2 Desc riptive statistics

Table4 shows descriptive statistics for the regression variables. As we estimate models

with different scalers, namely number of shares and market value of equity, two sets of

variables are shown in Panel A and B respectively. The second and third top rows of

Panel A show the mean (median) AGAAP earnings is 0.05 ( 0)per share and for equity


13/31


101

Journal

of

Contemporary Accounting Econo mics Vol4 ,No 2 (D ecember 2008)

89-119

it is 0.73 ( 0.23) per share. The mean and median

IFRS

earnings are similar to AGAAP

earnings but the equity mean and median are lower as Tables 2 and 3 suggest. The number

of

observations reported

in

Panel

B

is lower than in Panel

A

since valid 12-month returns

are needed, thereby excluding recently listed and thinly traded stocks. The sample size

reduces

to

922 from

1,065

for the returns model.

5.3Results

Panel

A

of Table

5

shows the price levels model results.

In

the cross sectional comparison

for the

AGAAP

regression, the coefficients for earnings and book value of equity are

Table 5

Relative Value Relevance

of AGAAP

and

IFRS

Earnings and Equity

Panel A : Price levels model

W a , , + a , E t ,+a , BV, , + E,,

Constant t-stat

E t-stat BV

t-stat

F-stat

Vuong Z-stat

(p-value)

Cross sectional comparison

N=

1,065)

IFRS

0.39 10.13** 1.95 9.01** 1.04 26.24** 0.62 878.15 (0.15)

AGAAP 0.34 9.52** 2.04 10.62** 1.01 31.82** 0.68 1,118.64 -1.44

Longitudinal comparison

(N =

972)

IFRS

0.39 7.51** 2.10 9.31** 1.19 24.63** 0.60 730.65 (0.01)

Panel B: Returns model

AGAAP 0.31 9.21** 2.44 11.44** 1.06 32.92** 0.73 1,336.35 -2.52

4 = a,, a J , ,

&,,

(2)

~ ~~

Constant t-stat E t-stat BV t-stat R2 F-stat VuongZ-stat

(p-value)

~~~~ ~

Cross sectional comparison

(N =

922)

AGAAP 0.24 7.84** 0.28 4.24**

IFRS 0.24 7.84** 0.34 4.28**

Longitudinal comparison (N

=

744)

AGAAP 0.21 7.24** 0.21 3.01**

0.02

17.94 0.03

0.02 18.28 (0.98)

0.01 9.03 0.30

IFRS

0.27 9.74** 0.18 3.55** 0.02 12.59 (0.62)

E

is annual AGAAPeamings,BV is book value of AGAAP equity at the end of the year. All independentvariables

are scaled by the number

of

shares at the end of the year in Panel A and by the market value

of

equity at the start

of the year in Panel B . Price is measured at the end of the last year that AGAAP

is

used in the cross sectional

comparison.Price is measured at the end of the respective financial year in the longitudinal comparison.Returns

are measured for the 12-month interval ending at the end of the last financial year that AGAAP is used in the

cross sectional comparison. Returns are measured for the 12-month interval ending at the end of the respective

financial year in the longitudinal comparison.**= significant at the 0.05 level (two-tailed test), * = significant

at the

0.10

level (two-tailed test)


14/31

102

John Goodwin,

Kamran

Ahmedand Richard Heaney

Journal of Contemporary Accounting Economics Vol4 ,No 2 December

2008)

89-119

2.04 and 1.01 respectively. For the IFRS regression the coefficients are 1.95 and 1.04

respectively. Both models are significant (p

ifrs on australian firms

Documents