3. Tsai-Yen.pdf

7/24/2019 3. Tsai-Yen.pdf

1/12

Global Economy and Finance JournalVol. 8. No. 1. March 2015 Issue. Pp. 3445

Are Special Items Relevant? A Case of Different Perspectives ofGAAP vs. IFRS

Ling-chi Cheng

*

and Tsai-yen Chung**

GAAP allows unusual or infrequent items to be reported as special items inorder to distinguish transitory earnings from permanent earnings. However,special items have been shown to hide permanent losses. IFRS prohibitsspecial items reporting to promote inter-company comparability. This studyexamines if special losses contain relevant information by using intra-industryinformation transfer methodology with a sample of 563 material assetimpairment losses reported as special items over the period from 2001 to2007, we find that such information is relevant but garbled due to the laxativereporting rules of GAAP. The major policy implication of this study is thatIFRS and GAAP should be more conciliatory by allowing special itemreporting, but at the same time requiring strict reporting and disclosure rules

to enhance the relevance and reliability of earnings information.

JEL Code: M41

1. Introduction

Accounting researchers have long studied the relation between earnings releases and securityprices. It is well known that the release of earnings information affects the security returnsbehavior of the announcing firms as well as non-announcing firms in the same industry(Freeman and Tse, 1992; Han and Wild, 1990).

This intra-industry information transfer is employed in this study to examine a contentious pointbetween GAAP and IFRS with respect to the treatment of special items in the income statement.IAS 1 specifically prohibits the presentation of any item of income and expense as extraordinaryitem either in the statement of comprehensive income or in the accompanying notes. The IASBconcludes that the so-called unusual in nature and/or infrequent in occurrence items areactually the results of the normal business risks faced by an entity and do not warrantpresentation as a separate component of the statement of comprehensive income. However,under GAAP, which is an income statement oriented reporting system for the valuation of thereporting entity, transitory income items should be separated from recurrent items to providecapital market with relevant information for valuation purposes.

The main critique leveled against the FASBs special items reporting rules is that they arelaxative and management routinely employs this provision for earnings management, i.e.,reporting operating losses as special items and one-time only gains as core earnings to biasinvestors perception of firms economic performance and share price (Cain et al. 2014). Thus,IASB insists that prohibition of special item reporting can result in greater comparability acrossfirms.

* Ling-chi Cheng, Department of Accounting, Chung Yuan Christian University, Taiwan, Email:[email protected]

** Tsai-yen Chung, College of Management, Yuan Ze University, Taiwan, Email:[email protected]
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]


2/12

Cheng & Chung

35

The contention between IFRS and GAAP centers on the opportunities for intra-period earningsmanagement. Although extant empirical evidence confirms the prevalence of intra-periodearnings management, the argument ignores the effect of special item reporting on themotivation for inter-periodearnings management. That is, when the losses are indeed transitoryand reporting them as special items can cast a less unfavorable light on the reporting entity,management may be more willing to include such losses in the income statement for the period

in which they occurred, rather than deferring to future periods. Although the affordedopportunities for earnings management cannot be completely eliminated, allowing managementto report special items actually renders more relevant information about special itemsannouncing firms earnings potential and this enhanced relevance may outweigh the loss ofreliability.

One of the sources of relevant accounting information is earnings of other firms in the sameindustry. Instead of concentrating on examining the stock price reactions of the special itemsreporting firms (SI firms), this study examines the relevance of special items by examining thecontemporaneous stock price behavior of the non-special items firms (NSI firms). Material assetimpairments are often industry phenomena, timely disclosure provides more relevant

information about the economic condition that may also affect other firms in the industry. Forexample, if timely inventory write-off had been provided by Lucent Technologies when dot combegan to unravel might have prompted analysts to provide timely investment advices andlessened the adverse effect of the bubble (Palepu et al. 2004). This information is not onlyvalue-relevant to Lucent, but also value-relevant to other firms in the same industry. Thus, whileconcentrating on intra-period earnings management of the reporting firms, accountingregulatory bodies should also consider the positive effect on the heightened incentive for timelyrecognition of losses in deciding whether to permit or forbid separately reporting special itemswhen management can substantiate that the item is indeed transitory.

This study examines the abnormal returns surrounding the announcement date of a sample of

563 firms that announced special item losses related to asset impairments equal to or largerthan 3% of total net assets in their annual earnings (SI firms) and contemporaneous pricemovements of all NSI firms in the same 4-digit SIC industry over the period from 2001 to 2007. Ifmaterial special items convey information about industry commonalities, the significance ofspecial items goes beyond simply about individual firms general business risk. If special itemsreporting practice motivates timely reporting of losses by management and provides investorswith more timely information about both the SI and NSI firms in the same industry, then thisreporting practice warrants special consideration by accounting rule-setting bodies.

The information transfer of special item announcements is tested using a standard event studymethodology. This study employs both the non-directional and the directional tests of the

security returns of SI and NSI firms with abnormal returns from the single-index market model.For the non-directional test, which is based on the contemporaneous return variability, theresults are generally consistent with the notion that special items provide value-relevantinformation about NSI firms in the same industry. The association indicates that special itemsare regarded by investors as value-relevant in reflecting information about the prevailingeconomic conditions in the industry.


3/12

Cheng & Chung

36

However, the results of directional tests do not produce statistically significant association ofstock price movements at 5% level between SI firms and NSI firms in the same industry. Theresult indicates that the information content of special items is garbled. An obvious reason, onecan argue, is due to the opportunities of earnings management. Although other explanationssuch as both the rival effects and the contagion effects are operative at the same time fordifferent firms in the same industry, our results imply that in order to increase the usefulness of

earnings information, accounting rule-setting bodies should permit special items reportingpractice but require stricter rules so that only truly transitory items can be reported as specialitems.

The remaining paper is organized as follows: Section 2 provides literature review; Section 3provides hypotheses development and methodology; Section 4 provides sample description;empirical results are provided in Section 5 and conclusions are provided in Section 6.

2. Literature Review

Earlier research on special items has examined market reactions at the release of special item

information and the whether managers use special items to manage earnings. Focusing onmaterial asset write-offs, Elliott and Shaw (1988) report significantly negative one- and two-daystock returns at the time of the announcements. Francis, Hanna, and Vincent (1996) concludethat the contemporaneous market reaction to special items depends on their nature. Forexample, their results show negative reactions to inventory write-offs (consistent with thewrite-offs conveying information about declines in economic circumstances) and positivereactions to special item charges (consistent with these items conveying information aboutimproving future prospects). Elliott and Hanna (1996) suggest at least two reasons to believethat the presence of special items in income makes it more difficult for users to determinerecurring earnings. First, firms may transfer current (or future) normal operating expenses intospecial items, thereby increasing current (or future) earnings before special items. Second,

write-offs may be associated with unusual and difficult to interpret economic circumstances.Kinney and Trezevant (1997) report that firms with large positive or large negative earningschanges are more likely to recognize negative special items, consistent with smoothing (for thepositive changes) and big bath (for negative changes) behaviors of inter-period transfers. Thesestudies provide comprehensive evidence that management uses special items reportingprovisions to engage in earnings management.

Recent studies focus on the nature of special items as an income construct and managementmotivation in reporting them. Riedl and Srinivasan (2010) report that special items receiving theincome statement presentation, as opposed to footnote disclosure, signals its lacking ofpersistence in future income. Cready et al. (2012) find that management uses restructuring

charges to signal real improvement of firms future earnings whereas asset write -offs do not.Their study focuses on taking big bath as inter-period transfer of loss. Our study argues thatspecial item reporting promotes timely recognition of losses and enhances relevance, ratherthan big bath behavior.

Cain et al. (2014) finds that only two-thirds of 13,174 special items over the period 1989-2011are actually transitory in nature. Cain et al. (2014) advocates permitting, but stricter reportingrules for special items since transitory items have different effects on firm valuation. Our study


4/12

Cheng & Chung

37

establishes the value relevance of special items by arguing that benefit of more flexiblereporting rules improves relevance by promoting timely recognition of losses.

Although the current study examines the difference between IFRS and GAAP in special itemsreporting rules, we do not intend to pit one against the other. Each has its merits and downsides.Van Tendeloo and Vanstraelen (2005) examine the difference in earnings management

behavior under IFRS after switching from GAAP. Presumably, high quality accounting standardsshould lessen the tendency of earnings management and IFRS is generally regarded as higherquality standards. However, Van Tendeloo and Vanstraelen (2005) did not find a significantdifference in earnings management under GAAP and IFRS. We argue that flexibility can betterpromote accounting quality than restrictions.

3. Hypotheses Formulation

An information transfer examined in this paper arises when the information releases of firm i(k,...., z) can be used to make inferences about the share price distribution of firmj,

),......,,,()( , zkjiijj PfPf

Where f (.) is a distribution function, iP is the share price of firm iand i ( j ) is firm i's (js)financial reporting system.

One motivation for this hypothesis arises from research documenting that the earnings of firmsare affected by (a) economy factors, (b) industry factors, and (c) firm specific factors. In thispaper, the concern is with (b). The earnings releases of other firms in the same industry are onesource of information on the impact of industry-wide trends for any single firm. This is especiallytrue in the case of material losses. Management may wish to disguise them as transitory lossesand report them as special items. These special item announcements contain information about

industry commonalities and constitute one possible source of information transfer arising fromthe earnings releases of firm i.

Two hypotheses regarding information transfer are tested in this paper. First, non-directionaltests of security returns are conducted to examine information transfers associated with specialitems announcements. The motivation for non-directional tests relates to the lack of a welldefined theory relating price reactions of special items and non-special items firms. In null form,the first hypothesis states that there is no abnormal return to the equity of non-announcing firmjassociated with the earnings release of announcing firm i, if iand jare in the same industry.Rejection of the null provides evidence that special items contain industry-relevant informationthat extends beyond the announcing firms.

Denoting ju as a residual return metric for firmjand i as the act of an earnings release byfirm i, this hypothesis may be stated formally as

0 : E( ) 0 ,j iH u

1 : E( ) 0 .j iH u


5/12

Cheng & Chung

38

If this null non-directional hypothesis is rejected, then a stronger hypothesis may be framedconcerning the direction of the abnormal returns experienced by special items and non-specialitems firms. The second hypothesis is twofold, i.e., the direction and magnitude of intra-industryinformation transfers are predicted to be related to the security return reaction of the specialitems firms. Specifically, the second null hypothesis examined is that the sign of the expectedabnormal return for NSI firmj, conditional on the abnormal returns of special items announcing

firm i, is unrelated to the sign of firm isabnormal return. The alternative hypothesis is that theconditional expected abnormal return for NSI firm j is in the same direction as the abnormalreturn of announcing firm i conditioned on its own earnings release. This hypothesis may bestated formally as

0 : E[ , E( ) 0 ] 0 ,j i i iH u u

1 : E[ , E( ) 0 ] 0 .j i i iH u u

4. Sample Selection

Our sample consists of firms that announced fourth quarter negative special items (Compustatitem #32) exceeding 3% of beginning fourth quarter assets from 2001-2007. Note that specialitems as reported by Compustat are not limited to those defined by GAAP, but also include itemsreported in the footnotes, but not shown separately on the income statement (Burgstahler et al.2002, p. 590). Our selection period starts in 2001 and terminates in 2007 to avoid the effectsfrom the capital market crash in 2008. From the initial sample, we eliminate special itemsrecorded after mergers, since these programs usually aim to eliminate redundancies rather thanto deal with troubled operations. We also exclude firms in Chapter 11 bankruptcy, as they aresubject to court oversight which may affect disclosure. We also impose the standard datarequirements of daily stock returns from the CRSP New York Stock Exchange (NYSE),

American Stock Exchange (AMEX), and NASDAQ files.

The sampling procedures yield the sample of 563 firms. We consider the four digits SIC codesfor each special items firm from CRSP. The average number of NSI firms for each SI firm is 10.5,ranging from 3 to 26.


6/12

Cheng & Chung

39

Table 1: Sample Selection of the Special Items

5. Methodology and Empirical Results

We employ the event study methodology to examine the information content of special items.The expected price is derived from the market model of the form suggested by Sharpe (1963,1964):

, , ,( )

i t i t i i M t u R R

where ,i tR , is the daily stock return for firm i on day t, ,M tR is the return on a value-weightedmarket portfolio for day t. The models parameters, i ,

i

are derived from ordinaryleast-squares regression using returns from day - 220 to day -21 (i.e., the estimation period),where day 0 is the special items announcement date.

5.1 Non-Directional Test Results

In non-directional test, we focus on abnormal return behavior (whether positive or negative) forthe special items and non-special items firms at the special items announcement date. Thisapproach is based on the concept that variability of abnormal returns of special item representsthe amount of unexpected (positive or negative) information about earnings containing special

item. For this purpose a standardized cumulative price variance (SCVR) metric, similar to theone derived from Foster (1981) is computed for firm i,

2

1, 2 1, 2

1

2

2

,{ } , ,{ }

t

i t t i t i t t

t t

SCVR u

Where 1, 2{ }t t denotes the event period from day 1t to day 2t , and 1, 22

,{ }i t t is an estimate of

var( 21

,

t

i tt tu

). The latter variance estimate is derived via

Panel A: Sample Selection Screening Results

Initial sample of special items firms 5472Special items and/or non-special items firms returns unavailablefrom CRSP

(1754)

Special items firms categorized in different four-digit SIC codeindustries within 250 trading days preceding the special items

(17)

Special items that do not meet asset impairment>3% of net assets (3138)Final sample 563

Panel B: Distribution of Sample by Year of Negative Special Items

Year No. of Firms Percent

2001 94 16.7%2002 73 13.0%2003 68 12.1%2004 82 14.6%2005 54 9.6%2006 113 20.0%

2007 79 14.0%Total 563 100.0%


7/12

Cheng & Chung

40

1, 2

2 2

,{ } 1 2 1( 2[1 2 ( 1) ]

i t t L L iL L S

where L equals ( 2 1 1t t ), l is the lag l autocorrelation coefficient of the residuals of thepricing model, and 2iS is the variance of the regressions error terms from the market modelsestimation period. This cumulative variance metric is derived for all special items andnon-special items firms.

Intra-industry information transfer is said to occur when the abnormal return variability of specialitem firms is related to that of the NSI firm at the time of special items announcements. The testis based on the spearman rank correlation ( s ) between SCVRs of SI and NSI firms at theportfolio level. In addition, regression analyses of non-special items firms SCVR on itscorresponding special items firms SCVRare conducted using both portfolio (p) and individual( i ) firm SCVRs,

SIxNSI

x SCVRSCVR *

wherex= {por i }

The expectations for the non-directional test, and consistent with the existence ofintra-industryinformation transfer,are s >0, >0, and >1.

The non-directional test results are presented in Table 2. Panel A of Table 2 presents the marketmodels SCVRreturn metrics for special items and non-special items firms. The SCVRin Panel

A.1 indicate a significant increase in price variability for special items firms at the special itemsannouncement date. The average SCVRfor the special items firms at {-1,0} is 3.244. The lastrow of Panel A1 presents the number of observations in each category at {-1,0}. Notice that ofthe 563 special items firms, 12% (67/563) have SCVRvalues less than 1.0 at {-1, 0}.

The abnormal return variability of non-special items firms is presented in Panel A.2 of Table 2.The average SCVRvalue at {-1.0} for non-special item firms is 1.172. A comparison of SCVRvalues of portfolios A through D for the special items and non-special items firms yieldsadditional insights. Both the price variability of SI and NSI firms at the time of special itemannouncements increased at announcement date. Although the return variability of non-specialitems firms is substantially less than the special items firms; however, the ranking order ofSCVRvalues of Portfolios A, B, and C for NSI firms is identical to that of the SI firm. The resultsindicate special items are value-relevant and therefore warrant separate reporting.

The analysis in Panel B of Table 2 extends that of panel A to conduct more powerful statisticaltests of return behavior. Panel B presents results from regressing the SCVRof NSI firms on the

SCVRof SI firms. For analysis at the portfolio level, the intercept ( r) is significantly differentfrom zero (at the 0.01 level), while the slope ( ) is significantly different from zero at 0.05 level, suggesting that non- special items firms exhibit above average price variabilityand that theirreaction is related (in magnitude) to the special items firms price reaction.For individual firmanalysis, the intercept ( r) and the slope ( ) are significantly greater than zero, suggesting thatnon-special items firms exhibit above average price variability, and that their reaction ispositively associated with (in magnitude) that of special items firms. Therefore, the results inTable 2 are consistent with following interpretation: (1) the return variability of NSI firms is


8/12

Cheng & Chung

41

consistent with information transfer of special items announcement; (2) at the portfolio level,special items firms with above (below) average price activity do yield significant above (below)average price activity for non-special items firms; and (3) the individual special items firms withabove (below) average price activity also yield significant above (below) average price activityfor non-special items firms in the same industry.

Table 2:Non-Directional Information Content and Transfer Testsa

aA *(**/***) designates statistical significance at the 0.10 (0.05/0.01) level, one-tailed tests.bSCVRs in panel A are from the market model. Panel A.2 is comprised of average SCVRs of non-special items

firms in the same cells as their corresponding special items firms in panel A.1. An approximate unit Normal variateis used to test the significance of SCVRexceeding 1.0.cSCVRs are computed for days {-1,0}. The numbers in parentheses are the t-statistics for the regression models

parameter estimates.

5.2 Directional Test Results

Evidence of non-directional information content and transfer motivate interest in strongerhypotheses concerning its directional nature. The directional tests of return behavior involvenonparametric Spearman rank correlation analyses and regression analyses. Two metrics aredeveloped to represent the unexpected information conveyed in special items announcements.The first metric utilizes the direction and magnitude of special item firms abnormal returns, the

cumulative abnormal return (CAR) for the special items firm i over the period 1t to 2t , is

2

1

21 ,,,~

t

tt

ti

SI

ttiCAR

Panel A: Non-directional Abnormal Return Behavior

Event timeGroup ASCVR:>5

Group BSCVR:2.5-5

Group CSCVR:1-2.5

Group DSCVR:0-1

Group A-DFull sample

Panel A. 1: Special Items Firms

(-5,-4) 0.875 0.934 1.153 0.868 0.991(-3,-2) 1.425 1.033 0.952 0.832 1.060(-1, 0) 6.934 3.426 1.754 0.794 3.244( 1, 2) 1.672 1.211 1.193 1.054 1.279( 3, 4) 0.925 1.006 0.867 0.841 0.922

Sample size (-1, 0) 114(20%) 187(33%) 195(35%) 67(12%) 563 (100%)

Panel A. 2: Non- special items Firms (-5,-4) 1.085 0.842 1.103 0.879 0.919(-3,-2) 1.221 1.234 1.106 0.834 0.936(-1, 0) 2.112 1.756 1.010 1.025 1.172( 1, 2) 1.862 0.914 0.833 0.802 0.891( 3, 4) 0.881 1.107 0.909 0.832 0.873

Sample size (-1, 0) 415(7%) 594(10%) 772(13%) 4155(70%) 5936(100%)

Panel B.: Regressions o f Non- special items firms SCVR on special items firms SCVRc

SIxNSI

x SCVRSCVR

n R F-ratio

Portfolio

(p)

20 0.96

(4.76)

0.62

(2.15)

0.23 4.61

Individual(i)

563 0.65(2.72)

0.57(2.95)

0.02 8.82


9/12

Cheng & Chung

42

Where,i tu is a directional measure of abnormal return activity.

The NSI firms reaction metric is similarly derived from its abnormal returns:

2

1

21 ,,,~

t

tt

tj

NSI

ttjCAR

The directional tests examine if CAR values of SI firms and those of NSI firms in the same

industry are correlated. Specifically, the special items firms are initially ranked in ascendingorder on the basis of CAR. Next, the firms are divided into twenty portfolios of an approximatelyequal number of firms, from which the average CAR for firms in each portfoliop, denoted SIpCAR ,is computed. Corresponding to each SI firms portfoliopis a portfolio comprised of NSI firms inthe same industries as the special items firms. The average cumulative abnormal returns forportfolio p of NSI firms portfolio vector is denoted NSIpCAR . The information transfer is thenmeasured by ),( NSIp

SI

ps CARCAR , where s , is the Spearman rank correlation coefficient and p=1,. . . ,20. In addition, the second metric is the percentage of positive CARduring the {-1,0}.

The second set of directional tests utilizes regressions of non-special items firms CAR onspecial items firms CARon a portfolio (p) and individual (i) firm basis. The specific regression is

SIxNSI

x CARCAR *

Where x= {por i}

The expectations for the directional tests using market model abnormal returns, and consistentwith positive industry cross-sectional covariance in returns, are s >0 and >0.

Table 3 presents the abnormal return behavior of the full sample of NSI firms categorized on theSI firms CAR. The market models abnormal returns for the full sample of non-special itemsfirms indicate no significant directional reaction (0.062%) at the time of the special items

announcement.

Panel B of Table 3 expands the descriptive analysis of directional return behavior in Panel A toincorporate more powerful statistical tests and to examine for any magnitude relation. Panel Adisplays market model abnormal returns for the SI firms (twenty portfolios, rank-ordered) andtheir corresponding NSI firms. The pattern of CARvalues of the non-special items firms tends tobe unrelated to those of the special items firms. Hence, this result is not consistent with thedirection and magnitude of a special items firms abnormal returns acting as a determinant ofthe direction and magnitude of the abnormal returns of other firms in its industry.


10/12

Cheng & Chung

43

Table 3: Directional Information Transfers Conditional on Special Items Firms CARa

Panel B.: Regressions o f Non- Special Items Firms CAR on Special Items Firms CAR

*NR RCAR CAR

n R F-ratio

Portfolio (p) 20 -0.01(0.02)

-0.22(1.09)

0.06 1.18

Individual (i) 563 -0.00(0.05)

0.12(1.31)

0.01 1.71

a

Cumulative abnormal returns are cumulated from days {-1,0}. A * (**/***) designates statistical significance at the0.10 (0.05/0.01) level, one-tailed tests.bPortfolios are comprised of nfirms and %POS is the percent of positive CAR among the non- special items firms.

Test of the significance of %POS is computed for the null hypothesis of %POS equal to the special items firms%POS sample proportionthe proportion is 0.50 for the market model abnormal returns.cThe Spearman rank correlation is computed between the special items (SI) and non- special items (NSI) firms

CAR portfolios.dParameter estimates and t-statistics (in parentheses) are presented for regressions using portfolio (p) and

individual (i) firm CAR.

Panel A: Abnorm al Return po rt fo l ios Condi t ional onSpecial items Firms CARb

SI firms NSI firms

Portfolio n CAR n CAR %POS

1 28 -0.198***

317 0.015 51

2 28 -0.148*** 206 -0.016 48

3 28 -0.125***

443 0.003 54*

4 28 -0.095***

208 -0.038* 49

5 28 -0.075***

361 -0.006 50

6 28 -0.054** 479 -0.059

** 46

**

7 28 -0.046** 197 0.043

* 52

8 28 -0.041** 442 0.039

* 53

*

9 29 -0.039** 302 0.031

* 49

10 29 -0.032** 168 -0.012 45

***

11 29 -0.025* 429 0.005 49

*

12 28 -0.017* 277 -0.038

* 46

*

13 28 -0.012 462 -0.015 50

14 28 -0.006 319 0.009 51

15 28 -0.022 276 -0.017 46*

16 28 0.016 206 -0.004 49

17 28 0.012 145 -0.001 51

18 28 0.026 205 -0.001 47

19 28 0.030 318 -0.004 47*

20 28 0.038 176 0.005 50

Spearman Rank Correlationcof SI Firms and NSI Firms CAR 0.042

Spearman Rank Correlation of SI Firms CAR and NSI Firms % Positive -0.209


11/12

Cheng & Chung

44

Panel A of Table 3 also includes the percent (%POS) of non-special items firms with a positiveCARover days {-1,0} for each portfolio. A binomial test of the significance of this percentage ofthe sample average is conducted for each portfolio. The binomial test results and the rankcorrelation indicate that the percent of positive CARs in the non-special items firm portfolios isunrelated to special items firm portfolios with largerCARvalues.

Panel B of Table 3 extends the analysis of Panel A to consider jointly any magnitude relation andany reaction to the act of the special items announcement on non-special item firms marketmodel returns. The regression results, based on portfolio and individual firms CAR, provide noevidence of a positive magnitude relation ( ) between special items and non-special item firmsabnormal returns. In portfolio level analysis, both the intercept ( r) and the slope ( ) are notsignificant. For individual firm analysis, the slope ( ) is positive as expected, however, themagnitude is insignificantly different from 0.

6. Conclusions

This paper investigates whether special item announcements affect the share prices of

non-special items firms in the same industry to establish if special items convey value-relevantearnings information. The sample is composed of 563 special items firms and 5936 firms in thesame industries during the period 2001-2007. We use both non-directional and directional testsof security return behavior to examine the association between special items firms and otherfirms in the same industries. Our results show that non-special items firms experiencesignificant abnormal returns variability associated with the earnings release of an SI firm in thesame industry. This result provides evidence that special items convey value relevantinformation beyond earnings for investors to evaluate the economic environment of the specialitems firms. The policy implication is that the IASBs strong position to prohibit special itemdisclosures in the income statement may not be warranted. Although the directional testsindicate that the information is not unambiguous, this lack of clear direction could be due to the

widely acknowledged earnings management opportunities due to the lax reporting rules. Asolution to this problem, as also advocated by Cain et al. (2014), can be stricter disclosure ruleswith respect to the nature and amount of special items. Accounting rule-setting bodies shouldadhere to the conceptual framework of providing relevant and reliable accounting information forinvestors and creditors to make economic decisions. This mandate requires rule setting bodiesto be more conciliatory rather than taking either or position in setting rules for special items.

References

Ali, A, J. Ronen and S. Li 1994, Discretionary Disclosures in Response to Intra-IndustryInformation Transfers, Journal of Accounting, Auditing and Finance, vol. 9, iss. 2, pp.

265-282.Atwood, TJ and Hong Xie 2010, The Market Mispricing of Special Items and Accruals: One

Anomaly or Two?, Review of Accounting and Finance,vol. 9, iss. 2, pp. 156-179.Burgstahler, D, J. Jiambalvo and T. Shevlin 2002, Do Stock Prices Fully Reflect the Implications

of Special Items for Future Earnings?, Journal of Accounting Research,vol. 40, no. 3, pp.585-612.


12/12

Cheng & Chung

45

Cain, CA, K. Kolev and S. McVay 2014, Qualifying Special Items, Unpublished Working Paper,pp. 1-48.AAA 2009 Financial Accounting and Reporting Section (FARS) Paper.Availableat SSRN:http://dx.doi.org/10.2139/ssrn.1267022

Elliott, J and JD. Hanna 1996, Repeated Accounting Write-offs and Information Content ofEarnings,Journal of Accounting Research, vol. 34, pp. 135-155.

Elliott, J and W. Shaw 1988, Write-offs as Accounting Procedures to Manage Perceptions,

Journalof Accounting Research, vol. 26, pp. 91-119.Foster, G 1981, Intra-industry Information Transfers Associated with Earnings Releases,

Journal of Accounting and Economics,vol. 3, iss. 3,pp. 201-232.Francis, J, JD. Hanna, and L. Vincent 1996, Causes and Effects of Discretionary Asset

Write-offs, Journal of Accounting Research, vol. 34, pp. 117-134.Freeman, R and S. Tse 1992, An Earnings Prediction Approach to Examining Intercompany

Information Transfers, Journal of Accounting and Economics,vol. 15, iss. 4,pp. 509-523.Han, JCY and JJ. Wild 1990, Unexpected Earnings and Intra industry Information Transfers:

Further Evidence, Journal of Accounting Research, vol. 28, no. 1, pp. 211-219.Han, JCY, JJ. Wild and K. Ramesh 1989, Managers' Earnings Forecasts and Intra industry

Information Transfers, Journal of Accounting and Economics,vol. 11, iss. 1,pp. 3-33.

Kinney, M and R. Trezevant 1997, The Use of Special Items to Manage Earnings andPerceptions, The Journal of Financial Statement Analysis, vol. 3, no. 1, pp. 45-53.

Palepu, KG, PM. Healy and VL. Bernard 2004, Business Analysis & Valuation, 3rd edn,Thomson South-Western, New York.

Pownall, G and G. Waymire 1989, Voluntary Disclosure Choice and Earnings InformationTransfer, Journal of Accounting Research,vol. 27, pp. 85-105.

Ramnath, S 2002, Investor and Analyst Reactions to Earnings Announcements of RelatedFirms: An Empirical Analysis, Journal of Accounting Research, vol. 40, iss. 5, pp.1351-1376.

Riedl, EJ and S. Srinivasan 2010, Signaling Firm Performance Through Financial StatementPresentation: Special Items, Contemporary Accounting Research, vol. 27, iss. 1, pp.

289-332.Tendeloo, BV and A. Vanstraelen 2005, Earnings Management under German GAAP versus

IFRS, European Accounting Review,vol. 14,iss. 1,pp. 155-150.
http://dx.doi.org/10.2139/ssrn.1267022http://dx.doi.org/10.2139/ssrn.1267022http://www.sciencedirect.com/science/journal/01654101/3/3http://www.sciencedirect.com/science/journal/01654101/15/4http://www.sciencedirect.com/science/journal/01654101/11/1http://www.sciencedirect.com/science/journal/01654101/11/1http://onlinelibrary.wiley.com/doi/10.1111/joar.2002.40.issue-5/issuetochttp://www.tandfonline.com/loi/rear20?open=14#vol_14http://www.tandfonline.com/toc/rear20/14/1http://www.tandfonline.com/toc/rear20/14/1http://www.tandfonline.com/loi/rear20?open=14#vol_14http://onlinelibrary.wiley.com/doi/10.1111/joar.2002.40.issue-5/issuetochttp://www.sciencedirect.com/science/journal/01654101/11/1http://www.sciencedirect.com/science/journal/01654101/15/4http://www.sciencedirect.com/science/journal/01654101/3/3http://dx.doi.org/10.2139/ssrn.1267022

Documents

3. Tsai-Yen.pdf