11.mahoney.et al 0104www.iiste.org call for_paper-130
Issues in Social and Environmental AccountingVol. 2, No. 1 June 2008Pp. 104-130 Corporate Social Performance, Financial Performance for Firms that Restate Earnings Lois Mahoney College of Business Eastern Michigan University, USA William LaGore College of Business Eastern Michigan University, USA Joseph A. Scazzero College of Business Eastern Michigan University, USAAbstractThis study examines corporate social performance (CSP) in firms that restate their financialstatements and, using a match pair design, compares their performance to firms that do notrestate their financial statements. Utilizing a randomized block design (two years prior to therestatement and two years after the restatement) for a sample of 44 U.S. firms, we found thatCSP Strengths, CSP Weaknesses, CSP People Strengths, and CSP People Weaknesses all in-creased after restatement though weaknesses increased at a greater rate than strengths. Addi-tionally, using panel data and a match pair design we found, we found that restating firms hada greater increase in CSP Strengths, CSP Weaknesses, CSP Product Strengths, CSP PeopleStrengths and a greater decrease in Total CSP People than non-restating firms after the restate-ment period. When comparing the relationships between CSP and financial performance (FP),we found that the positive relationship between ROA and CSP Strengths is greater for restate-ment firms than non-restating firms. In particular, we find that this positive relationship is aresult of the People dimension of CSP, in particular CSP People Strengths.Key Words: financial restatements, corporate social performance, financial performance,Lois Mahoney is Associate Professor of Accounting in the Department of Accounting and Finance College of Busi-ness Eastern Michigan University, USA, email: email@example.com. William LaGore is assistant professor,accounting and finance in the Department of Accounting and Finance College of Business Eastern Michigan Univer-sity, USA, email: firstname.lastname@example.org. Joseph A. Scazzeroa Professor of Decision Sciences in the Departmentof Accounting and Finance at Eastern Michigan University College of Business, USA, email: email@example.com
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 105CORPORATE SOCIAL PERFORM- of stakeholder confidence in financialANCE AND THE ISSUE OF FINAN- markets (Donoher et al., 2007) as well asCIAL RESTATEMENTS a firm’s corporate social performance (CSP). According to Carroll (1979),The quality of financial reporting has CSP considers a variety of factors, in-come under increased scrutiny in recent cluding discretionary responsibility toyears because of high-profile financial the community, economic responsibilityreporting failures, such as Enron and to investors and consumers, ethical re-WorldCom, and the significant increase sponsibilities to society and legal re-in the number of financial restatements. sponsibility to the government or theAn October 2002 General Accounting law. In this turbulent environment, theseOffice (GAO) report documents that the firms need to devise strategies that willnumber of financial restatements has enable them to survive and prosper inincreased 145 percent from 1997 to 2001 this environment in which stakeholdersand that publicly traded companies lost demand both financial performance (FP)billions in market capitalization in the and effective stakeholder responsivenessdays and months following a restatement (Johnson and Greening, 1999). Theseannouncement. The GAO report further firms may need to keep in mind CSP asconcludes that the increase in restate- they pursue superior performancements has negatively impacted investor through being responsive to the environ-confidence. For example, the GAO’s ment, maintaining product quality andOctober 4, 2002 letter to Senator Paul being responsive to the communities inSarbanes states the following: which it operates and the people it em- ploys (Turban and Greening, 1997). “The growing number of restate- ments and mounting questions Though research concerning the nature about certain corporate account- of the relationship between CSP and FP ing practices appear to have continues to be mixed (See Griffin and shaken investors’ confidence in Mahon, 1997; Roman et al., 1999), a our financial reporting system... number of findings indicate a positive empirical research studies and association (Worrell et al., 1991; Preston academic experts generally sug- and O’Bannon, 1997; Frooman, 1997; gest accounting issues have nega- Roman et al., 1999; Orlitzky and Benja- tively affected overall investor min, 2001; Murphy, 2002; Simpson and confidence and raised questions Kohers, 2002). Furthermore, most of about the integrity of U.S. mar- these findings are derived from compa- kets.” nies that are not experiencing financial reporting failures. The objectives of this paper are twofold: First, we address theLawsuits against firms resulted in nearly question of whether firms that restatea 1% loss in market value (Bhagat et al., financial statements have different levels1998), in which an estimated one-third of CSP than non-restating firms. Sec-of this loss is attributed to harmed firm ond, we address the questions onreputation (Karpoff and Lott, 1993). whether the relationship between CSPThe problem does not end with the pas- and FP is different between restatingsage of Sarbanes-Oxley Act (SOX) but firms and non-restating firms. This re-its continuation has implication for lack
106 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130search study will contribute to the ac- to financial reporting failures. Palmrosecounting research stream investigating et al. (2004) find a mean abnormal re-financial restatements, as well as the turn of -9.2% in the two-day windowethics research of CSP, and extends the (day 0, 1) around a restatement an-debate on the link between CSP and FP. nouncement, with more negative returnsThe remainder of the paper is organized for restatements involving fraud (-20%).as follows. First, we examine the back- Palmrose and Scholz (2004) find theground, theory and hypotheses. Second, negative market reaction is greater forwe explain our research methods and restatements of core earnings (i.e. pre-third, we present the results. The final tax earnings from primary operations)section includes our summary, discus- than for non-core earnings (i.e. all othersion and conclusions. earnings). Anderson and Yohn (2002) document the long-term economic con- sequences of restatements by findingBACKGROUND, THEORY AND average cumulative abnormal returns ofHYPOTHESES -7.97% for the period from three days before the restatement announcementFinancial Reporting Failures through three days after the restatement filing with the SEC.Financial reporting failures include bothfrauds and restatements. During the pe- There are also legal consequences to ariod of 1987-1996, the SEC found that a financial reporting failure. In the Palm-majority of frauds involved financial rose and Scholz (2004) study, 38 percentstatement fraud (Beasley et al., 1999). of the companies in their restatementThese frauds included sham sales, re- sample subsequently faced civil litiga-cording conditional sales as finalized tion. They found that companies withand recording revenues early. Thus, for restatements of core earnings (primarilythe purpose of this study, we examine revenue restatements) and pervasive re-only accounting restatements. Restate- statements (i.e. more than one account-ments are an admission that previously ing item restated) are more likely to beissued financial statements were not in subject to litigation.accordance with GAAP (Palmrose andScholz, 2004). Early research focused A financial reporting failure also dam-on characteristics of restating firms. For ages the reputation of the firm, auditors,example, Kinney and McDaniel (1989) management, and the board of directors.find restatement firms are smaller, less For example, Srinivasan (2005) foundprofitable, have higher debt, and are that outside board members experienceslower growing. DeFond and Jiambalvo significant reputational costs following(1991) find earnings overstatements are accounting restatements. Srinivasanmore likely for firms with diffuse own- finds significant turnover of board mem-ership and lower growth in earnings, and bers in the three years following the re-less likely for firms with audit commit- statement, including director turnovertees. for 48 percent of firms that restate earn- ings downward. The likelihood of direc-Recent studies document significant tor turnover increases if the board mem-negative economic consequences related ber is also on the audit committee. The
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 107study also finds outside directors lose reliable accounting numbers following apositions on other boards following a financial reporting failure.restatement. As for debt contracts, Sengupta (1998) suggests that quality of financial report-Another study, Desai et al. (2006), ex- ing is likely used by lenders in calculat-amined the reputational penalties to ing default risk. Sengupta (1998) foundmanagers of restating firms and found that firms with high disclosure qualitythat 60% of restating firms experience ratings from financial analysts aremanagement turnover in the two years charged a lower cost of debt, and thefollowing a restatement as compared importance of disclosure is greater whenwith 35 percent for a control sample. there is greater market uncertainty asAn audit firm’s reputation can be dam- measured by the variance of stock re-aged by a financial reporting failure, as turns. Thus, lenders likely demand aevidenced by the demise of Arthur An- higher risk premium following a report-dersen. Barton (2005) examines the de- ing failure in part due to the perceivedmand for auditor reputation by examin- decrease in quality of the accountinging the client defections from Arthur reports.Andersen. Barton (2005) finds firmsthat are more visible in the capital mar- The risk premium demanded by share-kets switched sooner to another Big 5 holders and debt holders also increasesauditor, as they were concerned about following a reporting failure because oftheir auditor’s reputation and the credi- the increased uncertainty about the fu-bility of their financial reporting. ture profitability and economic prospects of restatement firms. Palmrose et al.Accounting numbers used in contracts (2004) found a significant downward(e.g. compensation and debt contracts) revision in earnings forecasts followingmust be verifiable for the contract to be restatements and a significant increase inenforceable in court (Watts, 2003). analyst forecast dispersion (a proxy forBased on prior literature, it is reasonable uncertainty). Hribar and Jenkins (2004)to assume that a financial reporting fail- found accounting restatements lead toure leads to greater uncertainty about the decreases in expected future earnings.reliability and verifiability of the ac-counting numbers used in contracts. As In summary, restatements can have nu-a restatement casts doubt on the quality merous negative effects. These includeof the financial reports and increases the economic losses to investors; damage torisk to the contracting parties, sharehold- the reputations of the firm, auditors,ers and lenders will demand an increased management, and the board of directors;risk premium following a reporting fail- an increase in the cost of capital; and aure. For example, empirical studies find negative impact on future earningsthat frauds and accounting restatements power.lead to an increased cost of capital (e.g.,Dechow et al., 1996; Hribar and Jenkins, CSP and FP2004). Investors demand a higher rateof return to compensate for the per- Research on the relationship betweenceived riskiness of the firm due to less CSP and FP has resulted in positive (Wokutch and Spencer, 1987; McGuire
108 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130et al., 1988, 1990; Waddock and Graves, employee relations, environment, human1997; Simpson and Kohers, 2002; Or- rights, and product. The KLD indexlitzky and Benjamin, 2001; Mahoney ratings are based upon data gatheredand Roberts, 2007; Hill et al., 2007), from a broad range of sources; both in-negative (Waddock and Graves, 1997, ternal and external to the firm (see Wad-Preston and O’Bannon, 1997; Patten, dock and Graves, 1995 for further de-2002) and neutral results (Alexander and tails). Subsequently, this multidimen-Buchholz, 1978; Aupperle et al., 1985; sional index has been regarded as one ofUllman, 1985; Cochran and Wood, the best information sources available to1984; Shane and Spicer, 1983; Fauzi, researchers studying CSP (Hillman andforthcoming; Moore, 2001; Fauzi et al., Keim, 2001) and has been used in many2007). The negative view on the rela- subsequent studies (McGuire et al.,tionship between CSP and FP argues 2003; Hillman and Keim, 2001; Albin-that firms incur costs to improve social ger and Freeman, 2000, Greening andperformance and by doing so, they re- Turban, 2000; Mahoney and Roberts,duce profits and shareholder wealth. 2007; Mahoney et al., 2008; JohnsonThe positive view argues that better CSP and Greening, 1999).is viewed as positive by various stake-holders, leading to improved FP (Jones, Research Questions1995; Jones and Wicks, 1999). Thosewho support the neutral relationship ar- CSP: As discussed previously, there aregue that the direct relationship between significant negative economic, legal,CSP and FP does not exists due to the reputational, and contractual conse-complexity of the environment in which quences to a financial reporting failure.firms and society operate in (Mahoney A financial reporting failure is evidenceand Roberts, 2007) that previously issued accounting reports were incorrect, thus creating uncertaintyThe problem of measuring CSP is ar- about the credibility and verifiability ofgued by Waddock and Graves (1997) as financial reports after the reporting fail-the primary reason for the conflicting ure. In response to financial reportingresults found regarding the relationship failures, studies find firms take steps tobetween CSP and FP. Waddock and improve corporate governance mecha-Graves (1997) found a positive relation- nisms following a fraud or restatementship between CSP and FP when using an in order to restore credibility and trans-improved measurement of CSP, the parency in their financial reporting. ForKLD index. The KLD index provides example, Farber (2005) finds fraud firmsaccess to a wide range of independent, increase the number of audit committeeconsistently applied ratings of U.S. firms meetings and the number and percentageacross a number of important social per- of outside board members in the three-formance attributes that were determined year period following the fraud. LaGoreby a knowledgeable group of individuals (2008) finds restating firms significantlynot connected with the firms (Waddock increase the number of outside directorsand Graves, 1997). KLD evaluates each on the board, the number of audit com-company traded on the U. S. stock ex- mittee meetings, and the number of out-change over the dimensions of commu- side directors and financial experts onnity, corporate governance, diversity, the audit committee in the three-year
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 109period following a restatement an- H1: CSP (Total, Product and People)nouncement. These changes in corpo- before restatement is different than CSPrate governance may be mechanisms after restatement.that constrain management’s opportunis-tic behavior and lead to more transparent CSP and FP: As discussed before, em-reporting. However, it is unclear how pirical results concerning the nature ofthis improvement in corporate govern- the relationship, if any, between CSPance following a fraud or restatement and FP, continues to be mixed (See Grif-affects a firm’s CSP. fin and Mahon, 1997; Roman et al., 1999). Researchers have hypothesizedPrior research finds a positive relation- and have given rational theoretical justi-ship between disclosure level and CSP fication for negative, positive, and neu-(Gelb and Strawser, 2001). Mahoney et tral links between CSP and FP. Wad-al. (2008) examine CSP and executive dock and Graves (1997) argue that thecompensation before and after the Sar- fundamental reason for the uncertaintybanes-Oxley Act (SOX) and find that the between the CSP and FP relationship isimprovements in corporate governance the problem of measuring CSP. Hence,required by SOX may be resulting in Waddock and Graves (1997) used theincreased transparency regarding the KLD database as an improved measuremeasurement of CSP and an increase in of CSP and found a significant relation-accountability, as firms appear to be ship. Orlitzky (2008) found that there isstructuring compensation to promote an overall positive, but highly variableCSP. Gelb and Strawser (2001) also relationship between CSP and FP andfind that more extensive disclosures are noted that the large variability of find-provided by firms with higher CSP rat- ings in previous research is party due toings. Given that measures of CSP tend primary study artifacts. As studies findto rely on publicly available information, financial restatements negatively affectit may be that firms before the restate- firm performance and lead to increasedment would have been reluctant to make uncertainty about the future profitabilityfactors that are encapsulated in CSP and economic prospects of restatementweaknesses (bad news) available to the firms, it would be interesting to comparepublic. It follows that if improvements the association between CSP and FPin corporate governance following a re- between restating firms and non-statement encourage revelation and restating firms. Based upon these incon-transparency, the resulting increase in sistencies in prior research, it is unclearinformation available may influence how the negative effects of restatementsCSP in a negative direction. Further- on firm performance will impact CSP.more, following the restatement period, Thus, since we are unable to predict afirms may feel need to be more account- directional effect, the second researchable, thus influencing CSP in a positive question tested is as follows:direction. However, it would be difficultto theoretically determine the net direc- H2: The relationship between CSP andtional change in CSP as a result of re- FP is different for restating firms thanstatement. Therefore, the first research non-restating firms.question tested is:
110 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130METHODS earnings data are considered outliers with studentized residuals greater thanSample Selections the absolute value of three. Outliers are observations that are extreme or appearData on restating firms was obtained inconsistent with the remaining data.from the GAO-03-395R Financial State- This resulted in a final sample of 196ment Restatement Database for the pe- firms that had restated their financialriod of January 1, 1996 to June 30, 2002. statements. Missing CSP data for twoOf the initial sample of 919 restating years prior and two years after the re-firms, 40 firms were eliminated because statements reduced the final sample sizeno ticker symbol or CNUM could be to 44 firms. These 44 firms werefound. Ninety-three firms were deleted matched based upon SIC code to firmsbecause of multiple restatements. The that had not restated their financial state-initial collection of financial data found ments. Thirty-one companies werethat 153 firms were missing the required matched based upon the four-digit SICfinancial data. Furthermore, 174 firms code, five companies were matchedwere missing financial data in the post- based upon the last 3 digits of the SICrestatement period only, 200 firms were code and eight companies were matchedmissing financial data in the pre- based upon the last two digits of the SICrestatement period only, and 48 firms code. The final sample consisted fourwere missing financial data in both the years of data for 44 restating firms andpre- and post-restatement periods. The 44 non-restating firms, for a total num-missing data does not appear to be clus- ber of 88 firms with 352 observations.tered in either the pre- or post-restatement period. The number of The Modelfirms with missing Compustat data inthe pre-restatement period is comparable To test hypotheses 1, a randomizedto the post-restatement period, with 200 block design was used to determine theand 174 firms, respectively. Therefore, effect, if any, of restatement on the CSP.approximately 63 percent (575 firms) of To test hypotheses 2, panel data analysisthe initial restatement sample of 919 was used to examine the impact of re-firms did not have sufficient financial statement firms on the association be-data from Compustat to be included in tween the dependent variable CSP andthe final sample. This study requires the independent variable FP (ROA) withfinancial data for the two years prior to firm size, firm leverage and firm indus-and the two years following the restate- try as control variables. In order to cap-ment announcement year. A likely ex- ture omitted factors that may lead to aplanation for the loss of these firms is difference in CSP levels between thedue to the fact that many restating firms prestatement years and the postatementdeclare bankruptcy or are delisted fol- years, the indicator variable (as denotedlowing the restatements. This could po- by Post) is included as a separate inde-tentially lead to a survivorship bias, pendent control. Additionally, in orderwhich may prevent the results from gen- to capture the difference between restat-eralizing to the overall set of publicly ing firms and non-restating firms, thetraded firms. Finally, 15 firms were indicator variable (as denoted by Match)eliminated because their returns and is also included as a separate independ-
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 111ent variable. Two-factor interaction the prestatement years and the postate-terms are added to the model to allow ment years. The three-factor interactionthe effect of an independent variable on term ROA*Post*Match is added to thethe dependent variable to vary by the model to allow the ROA*Post interac-level of another independent variable. tion to differ between restating and non-For example, the interaction term restating firms. Hypotheses two is testedROA*Post allows the effect of ROA on through the following regression equa-the dependent variable CSP to differ for tion:CSPi,t+1 = b0 +b1ROAit + b2Matchit + b3Postit + b4ROA*Matchit + b5ROA*Postit + b6Match*Postsit + b7ROA*Post*Matchit (1) + b8Debt-to-Equityit + b9Assetsit + b10Industrykit results found regarding the characteris- i: firm tics of reporting firms, the quality of t: year their reporting, and the relationship be- k: 1-7 (number of SIC codes minus tween social performance and economic one) performance (Roberts, 1992; Gray et al., CSP = Corporate Social Perform- 1995). ance Score Value for Total, People, Product, Strengths and Shane and Spicer (1983) was one of the Weaknesses first published empirical studies to relyPost = 1 if one or two years after on externally produced ratings of CSP, restatement, 0 if otherwise using data developed by the U.S. Coun-Match = 0 if restatement firm, 1 oth- cil on Economic Priorities (CEP). They erwise argued that externally produced data wasROA =Return on Assets superior to voluntary disclosure whenDebt-to-Equity = Total Debt/Total Eq- performing cross-sectional studies, stat- uity ing:Industryk = 1 if industry k, 0 otherwise In the absence of mandated dis- closure and reporting standards, voluntary disclosures tend to beMeasures inconsistent and non-comparableDependent Variables from firm to firm, even in theMeasurement of CSP same industry. On the other hand, externally produced data (at leastAs prior research points out, there is no as produced by the CEP) washistory of systematic social reporting gathered using consistent proce-(Gray et al., 1995) and there are no gen- dures for collection and reportingerally accepted social reporting stan- across firms. Comparisons acrossdards (Wallage, 2000). Because of this, firms are thereby possible anddata for empirical research on CSP origi- potentially meaningful (p. 523).nates from voluntary disclosures byfirms or from external monitors. The Subsequent accounting studies alsoabsence of standardized reporting is at made use of CEP ratings (e.g., Cowen etleast partially responsible for the mixed al., 1987; Roberts, 1992).
112 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130In 1994, several U.S. researchers began et al., 2001). The KLD database hasto address the major problems in CSP been recognized as the best informationmeasurement by using the Kinder, available for researchers studying CSPLydenberg, Domini (KLD) database as a in the U.S. (Hillman and Keim, 2001).measurement of CSP. KLD rates over Therefore, we use KLD’s ratings of so-650 corporations traded on the U.S. cial performance to measure CSP.stock exchanges on various dimensionsconsidered important to social perform- Following previous research (Johnsonance. Because the KLD database was and Greening, 1999; Mahoney anddeveloped and maintained by an inde- Thorne, 2005), we use several differentpendent rating service that assessed CSP measurements of CSP that consider To-across a range of dimensions related to tal CSP, Total CSP Product, and Totalstakeholder concerns, researchers argued CSP People across the dimensions ofthat the KLD database brought a new strengths and weakness. CSP Strengthsand improved consistent measurement of are positive aspects of CSP; examplesCSP for United States companies include positive union relations, strong(Waddock and Graves, 1997). U.S. re- community giving, and environmentalsearch flourished with this new measure- planning. CSP Weaknesses are negativement assessment (Graves and Waddock, aspects of CSP; examples include safety1994; Waddock and Graves, 1997; Grif- problems, human rights violations, andfin and Mahon, 1997; Bendheim et al., environment fines. Figure 1 summarizes1998; Berman et al., 1999; Johnson and the different measures of CSP employedGreening, 1999; Greening and Turban, in this study.2000; Albinger and Freeman, 2000; Ruf Figure 1 Summary of CSP Measures Total CSP Variable CSP Strengths CSP Weaknesses Variable VariableTotal CSP Total CSP Total CSP Product Total CSP People(Community, Diversity, Em-ployee Relations, Environ-ment, International, Productand Business Practices andOther)Product Dimension (Product Total CSP Product CSP Product CSP Product and Business Practices and Strengths Weaknesses Environment)People Dimension Total CSP People CSP People CSP People (Community, Diversity and Strengths Weaknesses Employee Relations*Per Mahoney and Thorne (2005)
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 113Each company is given a Total CSP rat- ple. Total CSP People captures the con-ing by KLD along seven dimensions: tributions firms make to communitiescommunity, diversity, employee rela- through their hiring of women and mi-tions, environment, `international, prod- norities and their treatment of employ-uct and business practices, and other . ees. Executives may interpret the costsEach of these dimensions is given a of hiring minorities as unnecessarystrength rating and a weakness rating on short-term expenses; however, they maya scale of zero to two. A rating of 0 indi- recognize the long-term benefits of pro-cates no strengths or no weaknesses active employment policies when con-while a rating of 2 represents a major sidering the long-term avoidance ofstrength or a major weakness. CSP costly fines (Mahapatra, 1984). Further-Strengths are calculated by summing the more, signaling theory suggests that hir-strength ratings across all seven dimen- ing underrepresented groups sends asions for each company while CSP positive signal regarding a firm’s reputa-Weaknesses are calculated by summing tion and legitimacy (Turban and Green-the weakness ratings across all seven ing, 1997). Total CSP People is com-dimensions. Finally, Total CSP is calcu- posed of KLD’s dimensions of commu-lated by taking CSP Strengths and sub- nity, employee relations, and diversity.tracting CSP Weaknesses. Corporate governance would be ex-Our second measure of CSP is a sub- pected to have bearing and an associa-dimension of Total CSP: Total CSP tion on aspects or sub-dimensions ofProduct. Total CSP Product attempts to CSP that could be directly impacted bycapture the extent to which a firm is executives’ decisions while other sub-committed to quality products and prac- dimensions may be more impacted bytices sound environmental policies. For the general business or cultural contextexample, executives concerned with in which a firm operates. For example, aconsistent returns over time may likely firm’s diversity may be primarily im-avoid the imposition of costly environ- pacted by the labor pool that is available,mental fines (Johnson and Greening, while its product dimensions may be1999; Silverstein, 1994). Total CSP more easily impacted by executive’sProduct is comprised of KLD’s product attention to control and safety aspects inand business practices and environment product development. In fact, previousdimensions that relate to product and research has found differential associa-service quality and to the firm’s stance tions between some aspects of corporatetoward the natural environment. This governance and the people/product as-classification is consistent with ISO pects of CSP. For example, a positivestandards that require firms to establish a relationship for U.S. firms between topseries of management subsystems, stan- executive equity and the total productdards, and guidelines to ensure product dimension of CSP has been foundquality as well as safe and environmen- (Johnson and Greening, 1999), withouttally responsible practices (Uzumeri, comparable associations on the people1997). aspect of CSP.Our third measure of CSP is a sub- As discussed before, firms take steps todimension of Total CSP: Total CSP Peo- improve corporate governance mecha-
114 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130nisms following a fraud or restatement base. Industries are represented byin order to restore the credibility of their dummy variables and were broken downfinancial reports. In addition, it is ex- by four-digit Standard Industrial Classi-pected that corporate governance would fication (SIC) code per Graves and Wad-have bearing and an association on as- dock (1994).pects or sub-dimensions of CSP thatcould be directly impacted by execu- Panel Data Modelstives’ decisions. Thus it follows thatimprovements in corporate governance In summary, we investigate the behaviorfollowing a financial restatement may of CSP and its relation to FP by runningaffect certain aspects or sub-dimensions nine separate regressions using panelof CSP, particularly those that could be data—three regressions using CSP asdirectly affected by executive decisions. our dependent variable measure for To- tal CSP, Total CSP Product, and TotalIndependent Variables for Panel Data CSP People; three regressions usingAnalysis CSP Strengths for Total CSP Strengths, CSP Product Strengths, and CSP PeopleFollowing previous research, return on Strengths and three regressions usingassets (ROA) was used to measure a CSP Weaknesses for Total CSP Weak-firm’s FP (Waddock and Graves, 1997, nesses, CSP Product Weaknesses, andRoman et al., 1999, Mahoney and Rob- CSP Weaknesses, all with ROA as theerts, 2007; Fauzi, et al., 2007). Follow- independent variable.ing the works of prior research(Waddock and Graves, 1997; Mahoneyand Roberts, 2007), data on CSP was RESULTScollected for the year following the yearROA was reported to provide an oppor- Descriptive Statistics and Correlationtunity for capturing a lag between CSP Analysisand FP. Information on ROA was ob-tained from the Compustat database. Table 1 shows the means, standard de- viations, and correlations for our inde-Control Variables. Consistent with prior pendent, dependent, and control vari-research, we control for firm size, debt ables for the entire sample consisting oflevel and industry as previous research non-restating and restatement firms.noted that they may cause differences in The means for Total CSP, CSPFP (Waddock and Graves, 1997; Graves Strengths, and CSP Weaknesses are .15,and Waddock, 1994; Mahoney and Rob- 2.99, and 2.84 respectively. The meanserts, 2007). Consistent with prior re- for Total CSP Product, CSP Productsearch, total assets is used as a proxy for Strengths, and CSP Product Weaknessessize of the firm (Mahoney and Roberts, are -.61, .50, and 1.11 respectively. The2007; Graves and Waddock, 1994; Wad- means for Total CSP People, CSP Peo-dock and Graves, 1997) and debt-to- ple Strengths, and CSP People Weak-equity (Mahoney and Thorne, 2006) is nesses are 1.53, 2.39, and .86 respec-used to represent debt level. Informa- tively. The mean ROA is 6.07% and istion on total assets and debt-to-equity significantly positively correlated withare obtained from the Compustat data- Total CSP, Total CSP Product, and Total
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 115 Table 1 All Firms Pearson Correlation Matrix: Correlations with ROA, Control Financial Variables and Lagged CSPVariable Mean SD 1 2 3 4 5 6 7 8 9 10 11 1. Total CSP .15 3.132 2. CSP Strengths 2.99 2.537 .622** 3. CSP Weaknesses 2.84 2.523 -.616** .234** 4. Total CSP Product -.61 1.686 .707** .095 -.782** 5. CSP Product Strengths .50 .724 .378** .665** .200** .203** 6. CSP Product Weaknesses 1.11 1.694 -.542** .190** .863** -.908** .225** 7. Total CSP People 1.53 2.121 .762** .844** -.097 .155** .358** -.001 8. CSP People Strengths 2.39 2.010 .609** .952** .201** .050 .429** .134* .897** 9. CSP People Weaknesses .86 .944 -.415** .130* .646** -.241** .108* .286** -.338** .113**10. ROA 6.07% 6.78% .199** .074 -.173** .183** -.012 -.187** .189** .099 -.212**11. Debt-to-Equity 58.05% 18.86% -.148** .152** .336** -.361** .010 .363** .097 .193** .193** -.323**12. Assets $12,357 $16,104 -.120* .445** .596** -.410** .243** .512** .253** .441** .372** -.077 .241**
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 117CSP People and significantly negatively Weaknesses. Unlike non-restatingcorrelated with CSP Weaknesses, CSP firms, ROA for restatement firms is alsoProduct Weaknesses, and CSP People significantly positively related to TotalWeaknesses. The means for debt-to- CSP People and CSP People Strengthsequity is 58.05% and for assets are along with being significantly negatively$12,357 million. related to CSP Weaknesses and CSP People Weaknesses. Additionally, theTable 2 shows the means, standard de- mean debt-to-equity is 56.4 % and theviations, and correlations for our inde- mean assets are $13,713 million for non-pendent, dependent, and control vari- restating firms. Overall, restatementables for restatement firms only. The firms tend to have a higher level of Totalmeans for Total CSP, CSP Strengths, CSP, CSP Strengths, and CSP Peopleand CSP Weaknesses are .27, 3.11, and Strengths and CSP Weaknesses while2.84 respectively. The means for Total non-restatement firms have a higherCSP Product, CSP Product Strengths, level of CSP People Strengths and CSPand CSP Product Weaknesses are -.68, Product Weaknesses..49, and 1.18 respectively. The meansfor Total CSP People, CSP People Hypothesis 1Strengths, and CSP People Weaknessesare 1.69, 2.49, and .80 respectively. The To test hypothesis 1 a randomized blockmean ROA is 5.48% and is significantly design, equivalent to a paired t-test, waspositively correlated with Total CSP and used to determine the effect, if any, of aTotal CSP Product and significantly restatement on Total CSP, CSPnegatively correlated with CSP Product Strengths, and CSP Weaknesses; TotalWeaknesses. The mean debt-to-equity is CSP Product, CSP Product Strengths,59.7% and the mean assets are $11,000 and CSP Product Weaknesses; and Totalmillion. CSP People, CSP People Strengths, and CSP People Weaknesses. The depend-Table 3 shows the means, standard de- ent variable consisted of CSP scoresviations, and correlations for our inde- which were compared at different timependent, dependent, and control vari- periods, i.e., one year before and afterables for non-restating firms only. The restatement and two years before andmeans for Total CSP, CSP Strengths, after restatement. Table 4 summarizesand CSP Weaknesses are .03, 2.87, and the average Total, Strengths, and Weak-2.84 respectively. The means for Total nesses CSP scores for these time periodsCSP Product, CSP Product Strengths, and indicates which differences are sta-and CSP Product Weaknesses are -.55, tistically significant. Note that the aver-.51, and 1.05 respectively. The means age total for a score is equal to the dif-for Total CSP People, CSP People ference between the corresponding aver-Strengths, and CSP People Weaknesses age strength and average weakness.are 1.37, 2.28, and .91 respectively. Themean ROA is 6.64%. Similar to non- Most of the significant differences arerestating firms, ROA is significantly found by looking at two years beforepositively correlated with Total CSP and and two years after restatement. CSPTotal CSP Product and significantly Strengths and CSP Weaknesses signifi-negatively correlated with CSP Product cantly increased at p<.01 in the period
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 119 Table 4 Restating Firms Average CSP, Product, and People Scores for One and Two Year Time PeriodsDependent One Year Two YearsVariable Before After Difference Before After DifferenceTotal CSP 0.523 -0.136 0.659** 0.545 0.136 0.409CSP Strengths 3.068 3.114 -0.045 2.773 3.477 -0.705**CSP Weaknesses 2.545 3.250 -0.705** 2.227 3.341 -1.114**Total CSP Product -0.614 -0.841 0.227* -0.591 -0.682 0.091CSP Product 0.500 0.477 0.023 0.477 0.523 -0.046StrengthsCSP Product Weak- 1.114 1.318 -0.205 1.068 1.205 -0.136nessesTotal CSP People 1.773 1.636 0.136 1.705 1.636 0.068CSP People Strengths 2.455 2.523 -0.068 2.205 2.773 -0.568**CSP People Weak- 0.682 0.886 -0.205 0.500 1.136 -0.636**nesses*p<.05**p<.01following restatement though CSP change in CSP and test hypothesis two.Weaknesses increased by a greater In all equations, size, debt-to-equity ra-amount than CSP Strengths. This sig- tio, and industry were included as con-nificance appears to be driven by the trol variables. Consistent with prior lit-People dimensions of CSP as both CSP erature, a one-year lag between the FPPeople Strengths and CSP People Weak- variable and the dependent and controlnesses significantly increased at p<.01. variables was used.Also, CSP People Weaknesses increasedby a greater amount than CSP People Table 5 presents the results of our threeStrengths. When looking at one year panel data regressions that include Totalbefore and one year after restatement, CSP, CSP Strengths, and CSP Weaknesswe do find that Total CSP significantly as our dependent variable and ROA asdecreased and CSP Weaknesses signifi- our independent variable. For Totalcantly increased at p<.01. Additionally, CSP, similar to the results found in theTotal CSP Product significantly de- randomized block design, the Post vari-creased at p<.05. able was significantly negatively related at p<.05, indicating that CSP signifi-Hypothesis 2 cantly declined for all firms in the two years following the restatement period.Because we have cross-sectional and For the regression with CSP Strengths astime series data, we used panel data the independent variable, we found thatanalyses to further investigate the the Post variable was significantly posi-
120 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130tively related at p<.05 indicating that p<.01 indicating that the effect of ROACSP Strengths are higher after the re- on CSP Strengths also varies in the peri-statement period for all firms, which is ods prior and after restatement. For theconsistent with our results found in the regression using CSP Weaknesses as therandomized block design. Additionally, dependent variable, we found the Postwe found the interaction term of variable was significantly positively re-Match*Post was significantly negatively lated at p<.01, indicating that for allat p< .01. As shown in Figure 2, though firms the average CSP Weakness in-CSP Strengths increased for all firms creased in the period following the re-after the restatement period the increase statement. These results are also consis-was higher for restatement firms than tent with our findings in the randomizednon-restatement firms, suggesting that block design. Also the interaction termrestatement firms may be more account- of Match*Post was significantly nega-able after the period of restatement by tively related at p<.05. As shown in Fig-managing their CSP Strengths. The in- ure 3, restatement firms had a greaterteraction term of ROA*Match is signifi- increase in CSP Weaknesses than non-cantly negatively related at p<.01 indi- restatement firms. This is consistentcating that the effect of ROA on CSP with increased transparency followingStrengths is greater for restatement firms the restatement period as more negativethan non-restating firms, supporting hy- information concerning the firm is madepothesis 2. The interaction term of available.ROA*Post*Match was significantlypositively related to CSP Strengths at Table 6 presents the results of our three Table 5 Coefficient (Standard Error) of Panel Data Analysis for CSP Using a One Year Lag between the Dependent Variable and Independent VariablesDependent Total CSP CSP Strengths CSP WeaknessIndependent ROA -.015 .025 .009 .016 .025 .018 Match .293 .669 .483 .498 .184 .431 Post -.653 .272* .396 .180* 1.053 .195** ROA*Match -.074 .039 -.084 .026** -.009 .028 ROA*Post .024 .034 -.012 .023 -.038 .025 Match*Post -.426 .400 -1.124 .266** -.732 .286* ROA*Post*Match .073 .046 .103 .031** .032 .033Control Debt-to-Equity -.184 1.169 .609 .804 1.009 .802 Assets -.001 .000 .001 .000** .001 .000**R2 .151 .284 .464Wald chi-square 40.3** 66.8** 133.46**Panel data model typeNumber of Firms 88 88 88Number of Observations 352 352 352*p<.05**p<.01
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 121 Figure 2 CSP Strengths 3.4 3.3 Restatement CSP Strengths 3.2 3.1 Firms 3.0 2.9 Non- 2.8 Restatement 2.7 Firms 2.6 2.5 Prior to After Restatement Restatement Time Period Figure 3 CSP Weaknesses 3.4C S P W eakn esses 3.2 3.0 Restatement Firms 2.8 2.6 Non-Restatement 2.4 Firms 2.2 2.0 Prior to After Restatement Restatement Time Period
122 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130panel data regressions that include Total Product Strengths as the dependent vari-CSP Product, CSP Product Strengths, able, we found that the interaction termand CSP Product Weakness as our de- of Match*Post was significantly nega-pendent variable and ROA as our inde- tively related at p<.05. As shown in Fig-pendent variable. For Total CSP Prod- ure 4, restatement firms showed a slightuct, we found no significant relation- increase in CSP Product Strengths in theships. For the regression with CSP period following restatement while non- Table 6 Coefficient (Standard Error) of Panel Data Analysis for CSP Using a One Year Lag between the Dependent Variable and Independent VariablesDependent Total CSP Product CSP Product CSP Product Weak- Strengths nessIndependent ROA -.009 .011 -.004 .005 .005 .009 Match .219 .319 .066 .152 -.155 .297 Post -.181 .118 .026 .057 .206 .104* ROA*Match -.003 .017 -.000 .008 .004 .015 ROA*Post .010 .015 -.005 .007 -.015 .013 Match*Post .057 .174 -.168 .083* -.218 .153 ROA*Post*Match -.004 .020 .009 .010 .013 .018Control Debt-to-Equity -.306 .522 -.132 .250 .124 .467 Assets -.001 .000** .001 .000 .001 .000**R2 .3363 .177 .407Wald chi-square 62.72** 26.26** 74.52**Panel data model typeNumber of Firms 88 88 88Number of Observations 352 352 352*p<.05**p<.01restatement firms showed a decrease. CSP People, CSP People Strengths, andAgain, suggesting that restatement firms CSP People Weakness as our dependentmay be more accountable after the pe- variable and ROA as our independentriod of restatement and managing their variable. For our regression with TotalCSP Product Strengths. For the regres- CSP People, the interaction term ofsion using CSP Product Weaknesses as ROA*Match is significantly negativelythe dependent variable, only the Post related at p<.01 indicating that the effectterm was significantly positively related of ROA on Total CSP People is greaterat p<.05 indicating that CSP Product for restatement firms than non-restatingWeaknesses increased for all firms in the firms, supporting hypothesis 2. The in-period following the restatement. teraction term of ROA*Post*Match is also significantly positively related toTable 7 presents the results of our three Total CSP People at p<.01 indicatingpanel data regressions that include Total that the effect of ROA on CSP Strengths
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 123 Figure 4 CSP Product Strengths C SP P ro d u c t Stre n g th s 0.60 0.55 0.50 Restatement Firms 0.45 Non-Restatement 0.40 Firms 0.35 0.30 Prior to After Restatement Restatement Time Periodalso varies in the periods prior and after Strengths also varies in the periods priorrestatement. The interaction term of and after restatement. The interactionMatch*Post is significantly negatively term of Match*Post is significantlyrelated at p<.05. Per figure 5, restate- negatively related at p<.05. Per figure 6,ment firms had a bigger decrease in To- restatement firms had a bigger increasetal CSP People after restatement than in CSP People Strengths after restate-non-restatement firms. For the regres- ment than non-restatement firms, againsion using CSP People Strengths as the suggesting the restatement firms may bedependent variable, we found that the more accountable in the period follow-Post variable is significantly positively ing restatement by focusing in on CSPrelated at p>.05 indicating that for all strengths. For our CSP People Weak-firms CSP People Strengths significantly nesses regression, the only significantincreased in the two years following the variable that we found was the Post vari-restatement period. We also found that able at p<.01, indicating that for allthe interaction term of ROA*Match is firms CSP People Weaknesses signifi-significantly negatively related at p<.01, cantly increased in the two years follow-indicating that the effect of ROA on ing the restatement period.CSP People Strengths is greater for re-statement firms than non-restating firms,supporting hypothesis 2. The interaction SUMMARY AND DISCUSSIONterm of ROA*Post*Match is also sig-nificantly positively related to CSP Peo- This study was undertaken to investigateple Strengths at p<.01, indicating that CSP in restatement firms along with in-the effect of ROA on CSP People vestigating the relationship of CSP to FP
124 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 Table 7 Coefficient (Standard Error) of Panel Data Analysis for CSP Using a One Year Lag between the Dependent Variable and Independent VariablesDependent Total CSP People CSP People CSP People Weak- Strengths nessIndependent ROA .021 .018 .015 .014 -.006 .011 Match .268 .446 .431 .394 .214 .202 Post -.042 .197 .324 .153* .378 .128** ROA*Match -.093 .028** -.084 .022** .001 .018 ROA*Post -.015 .025 -.007 .019 .005 .016 Match*Post -.715 .290* -.881 .225** -.212 .188 OA*Post*Match .107 .034** .093 .026** -.010 .022Control Debt-to-Equity .213 .821 .793 .668 .576 .419 Assets .001 .000* .001 .000** .001 .000** 2R .166 .283 .203Wald chi-square 41.06** 80.66* 54.03**Panel data model typeNumber of Firms 88 88 88Number of Observa-tions 352 352 352*p<.05**p<.01for these same firms. Consistent with We also compared restatement firmsprior research on accountability and dis- with matched non-restating firms in ourclosure (Mahoney et al., 2008), we panel data analysis. We found that CSPfound Total CSP after restatement of Strengths, CSP Weaknesses, and CSPearnings was significantly lower than the People Strengths for restatement firmsaverage Total CSP before restatement. showed a greater increase than non-In particular, even though CSP Strengths restatement firms. For Total CSP Peo-increased, it was offset by a greater in- ple, we found that restatement firmscrease in CSP Weaknesses. This in- showed a greater decrease than non-crease in strengths may be due to the restatement firms. For CSP Productefforts by the firms to be accountable Strengths, we found that while restate-and improve the reputation of the firm. ment firms increased slightly, non-However, this may have been offset by restating firms showed a significant de-the negative impact of transparency sur- crease. These findings are consistentrounding financial restatement. These with prior research on reporting failurefindings support hypothesis 1 for Total that show that restating firms take stepsCSP which differs before and after re- to improve corporate governance mecha-statement. nisms following restatement in order to restore credibility and transparency
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 125(Farber, 2005; LaGore, 2008) and pro- tiques of KLD’s perspective on CSPvide addition support for hypothesis 1. would aid in the development of this research stream.We also find support for hypothesis 2,since a stronger positive relationship The sample selection bias is also a po-exists between ROA on CSP Strengths tential alternative explanation of the re-for restatement firms than non-restating sults. There are some possible selectionfirms. In particular, we find that this biases in our final sample of restatementeffect is a result of the People dimension firms since the research design requiresof CSP with significant relationships for each sample firm to have data for a con-Total CSP People and CSP People secutive 5-year period, the two yearsStrengths while no relationship was before and after the restatement an-found for any dimension of CSP Prod- nouncement. Thus, the final sampleuct. These results provide further sup- tends to include surviving and largerport for the previous literature on the firms that may be perceived as more re-positive relationship between CSP and liable. Therefore, the external validityFP and that CSP and FP may be mutu- of the study may be in question as theally reinforcing organizational activities results may not generalize to the overall(Orlitzky, 2008). population of publicly traded companies. On the other hand, larger firms receiveLike all research, ours has limitations more media coverage and regulatoryassociated with the measures, methodol- attention than smaller firms and there-ogy and sample size. The use of KLD fore may be under more pressure toratings to measure CSP are questionable change financial reporting and corporate(Chatterji and Levine, 2006; Chatterji, et social performance following a restate-al., forthcoming; Orlitzky and Swanson, ment in order to restore the public’s trust2008; Porter and Kramer, 2006) since in their financial reporting. The resultsthey are determined by an independent of this analysis are encouraging becausefirm and are the result of Kinder, Lyden- the prospect of a positive CSP and FPberg, Domini Research & Analytics’ ownership links means that even restate-definition and evaluations of CSP. Pre- ment firms can be socially responsiblevious research has found that while KLD and financially successful following theweakness ratings are a good summary of period of restatement.past environmental performance, KLDstrengths do not accurately predict pollu-tion levels or compliance violations REFERENCES(Chatterji et al., forthcoming). Researchhas also found that KLD is not optimally Albinger, H. S. & Freeman, S. J. (2000)using publicly available data (Chatterji “Corporate Social Performanceet al., forthcoming). Furthermore, the and Attractiveness as an Em-equal weighting and content of each di- ployer to different job seekingmension of CSP is another limitation populations”, Journal of Business(Chatterji and Levine, 2006). Future Ethics, Vol. 28, No. 3, pp. 243-research on the investigation of the con- 253.struction validity of KLD, the impact of Alexander, G. & Buchholz, R. (1978)equal weighting of dimension and cri- “Corporate Social Responsibility
126 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 and Stock Market Performance”, Performance”, Academy of Man- Academy of Management Journal, agement Review, Vol. 4, No. 4, Vol. 21, No. 3, pp. 479-486. pp. 497-505.Anderson, K. & Yohn, T. (2002) “The Chatterji, A.K., & Levine, D. (2006) Effect of 10-K Restatements on “Breaking Down the Wall of Firm Value, Information Asym- Codes: Evaluating Non-Financial metries, and Investors’ Reliance Performance Measurement”. Cali- on Earnings”, Working Paper, fornia Management Review, 48 Georgetown University. (2): 29-51.Aupperle, K.E., Carroll, A.B. & Hat- Chatterji, A. K., Levine, D. & M.W. field, J.D. (1985) “An Empirical Toffel. (Forthcoming). “How Examination of the Relationship Well do Social Ratings Actually between Corporate Social Re- Measure Corporate Social Re- sponsibility and Profitability”, sponsibility?” Journal of Econom- Academy of Management Journal, ics and Management Strategy. Vol. 28, No. 2, pp. 446-463. Cochran, F.L. & Wood, R.A. (1984)Barton, J. (2005) “Who Cares About “Corporate Social Responsibility Auditor Reputation?” Contempo- and Financial Performance”, rary Accounting Research, Vol. Academy of Management Journal, 22, No. 3, pp. 549-586. Vol. 27, No. 1, pp. 42-56.Beasley, M.S., Carcello, J.V. & Herman- Cowen, S., Ferreri, L. & Parker, L. son, D.R. (1999) “Just say ‘No’”, (1987) “The Impact of Corporate Strategic Finance, Vol. 80, No. Characteristics on Social Respon- 11, pp. 52-57. sibility Disclosure: A TypologyBendheim, C.L., Waddock, S.A. & and Frequency-Based Analysis”, Graves, S.B. (1998) “Determining Accounting, Organizations & So- Best Practice in Corporate- ciety, Vol. 12, No. 2, pp. 111-122. Stakeholder Relations Using Data Dechow, P., Sloan, R. & Sweeney, A. Envelopment Analysis”, Business (1996) “Causes and Consequences and Society, Vol. 37, No. 3, pp. of Earnings Manipulations: An 305-338. Analysis of Firms Subject to En-Berman, S.L., Wicks, A.C., Kotha, S. & forcement Actions by the SEC” Jones, T.M. (1999) “Does Stake- Contemporary Accounting Re- holder Orientation Matter? The search, Vol. 13, No. 1, pp. 1-36. Relationship between Stakeholder DeFond, M. & Jiambalvo, J. (1991) Management Models and Firm “Incidence and Circumstances of Financial Performance”, Academy Accounting Errors”, The Account- of Management Journal, Vol. 42, ing Review, Vol. 66, No. 3, pp. No. 5, pp. 488-506. 643-655.Bhagat, S., Bizjak, J. & Coles, J.L. Desai, H., Hogan, C. & Wilkins, M. (1998) “The Shareholder Wealth (2006) “The Reputational Penalty Implications of Corporate Law- for Aggressive Accounting: Earn- suits”, Financial Management, ings Restatements and Manage- Vol. 27, No. 4, pp. 5-27. ment Turnover”, The AccountingCarroll, A. (1979)“A Three-Dimensional Review, Vol. 81, No. 1, pp. 83- Conceptual Model of Corporate 112.
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 127Donoher, W.J., Reed, R. & Storrud- 37, No. 4, pp. 1034-1046. Barnes, S.F. (2007) “Incentive Gray, R., Kouhy, R. & Lavers, S. (1995) Alignment, Control and the Issue “Corporate Social and Environ- of Mislead Financial Disclosures’. mental Reporting: A Review of Journal of Management, 33 the Literature and a Longitudinal (4):547-569. Study of UK Disclosure”, Ac-Farber, D. (2005) “Restoring Trust after counting, Auditing & Account- Fraud: Does Corporate Govern- ability Journal, Vol. 8, No. 2, pp. ance Matter?” The Accounting 47-77. Review, Vol. 80, No. 2, pp. 539- Greening, D.W. & Turban, D. B. (2000) 561. “Corporate Social Performance asFauzi, H., Mahoney, L. (forthcoming). a Competitive Advantage in At- “Identifying and Analyzing the tracting a Quality Workforce”, Level of Practices of Company’s Business and Society, Vol. 39, No. Social Responsibility in Improv- 3, pp. 254-280. ing Financial Performance”, Jour- Griffin, J.J. & Mahon, J.F. (1997) “The nal of Business and Management. Corporate Social Performance and________, ________, & Rahman, A. A. Corporate Financial Performance (2007) “The Link between Corpo- Debate” Business and Society, rate Social Performance and Fi- Vol. 36, No. 1, pp. 5-31. nance Performance: Evidence Hill, R.P., Ainscough, T., Shank, T. & from Indonesian Companies”, Manullang, D. (2007) “Corporate Issues in Social and Environ- Social Responsibility and Socially mental Accounting, Vol. 1, No. 1, Responsible Investing: A Global pp. 149-159. Perspective”, Journal of BusinessFrooman, J. (1997) “Socially Irresponsi- Ethics, Vol. 70, No. 2, pp. 165– ble and Illegal Behavior and 174. Shareholder Wealth”, Business Hillman, A.J. & Keim, G.D. (2001) and Society, Vol. 36, No. 3, pp. “Shareholder Value, Stakeholder 221-249. Management, and Social Issues:Gelb, D. S. & Strawser, J. S. (2001) What’s the Bottom Line?” Strate- “Corporate Social Responsibility gic Management Journal, Vol. 22, and Financial Disclosures: An No. 2, pp. 125-139. Alternative Explanation for In- Hribar, P. & Jenkins, N. (2004) “The creased Disclosure”, Journal of Effect of Accounting Restate- Business Ethics, Vol. 33, pp. 1-13. ments on Earnings Revisions andGeneral Accounting Office (2002) the Estimated Cost of Capital”, “Financial Statement Restate- Review of Accounting Studies, ments: Trends, Market Impacts, Vol. 9, pp. 337-356. Regulatory Responses, and Re- Johnson, R.A. & Greening, D.W. (1999) maining Challenges”. GAO-03- “The Effects of Corporate Gov- 138. ernance and Institutional Owner-Graves, S.B., & Waddock, S.A. (1994) ship Types on Corporate Social “Institutional Owners and Corpo- Performance”, Academy of Man- rate Social Performance”, Acad- agement Journal, Vol. 42, No. 5, emy of Management Journal, Vol. pp. 564-576.
128 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130Jones, T.M. (1995) “Instrumental Stake- rate Social Responsibility: A Ca- holder Theory: A Synthesis of nadian Investigation”, Journal of Ethics and Economics”, Academy Business Ethics”, Vol. 69, No. 2, of Management Review, Vol. 20, pp. 149-162. No. 2, pp. 404-437. _________, Roush, P. & Thorne, L._______ & Wicks, A.C. (1999) (2008) “An Empirical Investiga- “Convergent Stakeholder The- tion of the Relationship of Corpo- ory”, Academy of Management rate Social Responsibility and Ex- Review, Vol. 24, No. 2, pp. 206- ecutive Compensation: U.S. ver- 221. sus Canada”, Working Paper.Karpoff, J.M. & Lott, J.R. (1993) “The McGuire, J.B., Schneeweis, T. & Reputational Penalty Firms Bear Branch, B. (1990) “Perceptions for Committing Criminal Fraud”, of Firm Quality: A Cause or Re- Journal of Law and Economics, sult of Firm Performance”, Jour- Vol. 36, pp. 757-802. nal of Management, Vol. 16, No.Kinney, W. & McDaniel, L. (1989) 1, pp. 167-180. “Characteristics of Firms Correct- __________, ________ & Sundgren, A. ing Previously Reported Quarterly (1988) “Corporate Social Respon- Earnings”, Journal of Accounting sibility and Firm Financial Per- and Economics, Vol. 11, No. 1, formance”, Academy of Manage- pp. 71-93. ment Journal, Vol. 31, No. 4, pp.LaGore, W.D. (2008) “Conditional and 854-872. Unconditional Conservatism Fol- __________, Dow, S. & Argheyd, K. lowing a Financial Reporting Fail- (2003) “CEO Incentives and Cor- ure: An Empirical Study”, Disser- porate Social Performance”, Jour- tation, Florida State University. nal of Business Ethics, Vol. 45,Mahapatra, S. (1984) “Investor Reaction pp. 341-359. to a Corporate Social Account- Moore, G. (2001) “Corporate Social and ing”, Journal of Business, Fi- Financial Performance: An Inves- nance & Accounting, Vol. 11, No. tigation in the UK Supermarket 1, pp. 29-40. Industry”, Journal of BusinessMahoney, L.S. & Roberts, R. (2007) Ethics, Vol. 34, No. 3/4, pp. 299- “Corporate Social Performance, 315. Financial Performance and Insti- Murphy, E. (2002) “Best Corporate Citi- tutional Ownership in Canadian zens Have Better Financial Per- Firms”, Accounting Forum, Vol. formance”, Strategic Finance, 31, No. 3, pp. 233-253. Vol. 83, No. 7, pp. 20-21._________ & Thorne, L. (2005) Orlitzky, M. (2008) “Corporate Social “Corporate Social Responsibility Performance and Financial per- and Long-term Compensation: formance: A Research Synthesis” Evidence from Canada”, Journal In A. Crane, A. McWilliams, D. of Business Ethics, Vol. 57, No. 3, Matten, J. Moon & D. Siegel pp. 241-253. (Eds.), The Oxford Handbook of_________ & _______ (2006) “An Ex- CSR (pp. 113-134). Oxford, UK: amination of the Structure of Ex- Oxford University Press. ecutive Compensation and Corpo- _________, & Benjamin, J.D. (2001)
L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 129 “Corporate Social Performance (1999) “The Relationship between and Firm Risk: A Meta-Analytic Social and Financial Perform- Review”, Business and Society, ance”, Business and Society, Vol. Vol. 40, No. 4, pp. 369-396. 38, No. 1, pp. 109-125._________, & Swanson, D.L. (2008) Ruf, B.M., Muralidhar, K., Brown, Toward integrative corporate citi- R.M., Janney, J.J. & Paul, K. zenship: Research advances in (2001) “An Empirical Investiga- corporate social performance. tion of the Relationship between London: Palgrave Macmillan. Change in Corporate Social Per-Palmrose, Z., Richardson, V. & Scholz, formance and Financial Perform- S. (2004) “Determinants of Mar- ance: A Stakeholder Theory Per- ket Reactions to Restatement An- spective”, Journal of Business nouncements”, Journal of Ac- Ethics, Vol. 32, No. 2, pp. 143- counting and Economics, Vol. 37, 156. No. 1, pp. 59-89. Sengupta, P. (1998) “Corporate Disclo-_________ & Scholz, S. (2004) “The sure Quality and the Cost of Debt. Circumstances and Legal Conse- The Accounting Review, Vol. 73, quences of non-GAAP Reporting: No. 4, pp. 459-474. Evidence from Restatements”, Shane, P.B. & Spicer, B.H. (1983) Contemporary Accounting Re- “Market Response to Environ- search, Vol. 21, No. 1, pp. 139- mental Information Produced Out- 180. side the Firm”, The AccountingPatten, D.M. (2002) “The Relation be- Review, Vol. 58, No. 3, pp. 521- tween Environmental Perform- 536. ance and Environmental Disclo- Silverstein, K. (1994) “Pension Funds sure: A Research Note”, Account- Increase Presence in Board- ing Organizations and Society, rooms”, Pension World, Vol. 30, Vol. 27, No. 8, pp. 763. No. 5, pp. 4.Porter, M. E., & Kramer, M.R. (2006) Simpson, W.G. & Kohers, T. (2002) “Strategy and Society: The Link “The Link between Corporate So- between Competitive Advantage cial and Financial Performance: and Corporate Social Responsibil- Evidence from the Banking Indus- ity”, Harvard Business Review, try”, Journal of Business Ethics, Vol. 84, No. 12, pp. 78-92. Vol. 35, No. 2, pp. 97-109.Preston, L.E. & O’Bannon, D.P. (1997) Srinivasan, S. (2005) “Consequences of “The Corporate Social-Financial Financial Reporting Failure for Performance Relationship”, Busi- Outside Directors: Evidence from ness and Society, Vol. 36, No. 4, Accounting Restatements and Au- No. 419-429. dit Committee Members” JournalRoberts, R.W. (1992) “Determinants of of Accounting Research, Vol. 43, Corporate Social Responsibility No. 2, pp. 291-334. Disclosure: An Application of Turban, D.B. & Greening, D.W. (1997) Stakeholder Theory”, Accounting, “Corporate Social Performance Organizations & Society, Vol. 17, and Organizational Attractiveness No. 6, pp. 595-612. to Prospective Employees”, Acad-Roman, R.M., Hayibor, S. & Agle, B.R. emy of Management Journal, Vol.
130 L. Mahoney, W. LaGore, J. A. Scazzaro / Issues in Social and Environmental Accounting 1 (2008) 104-130 40, No. 3, pp. 658-672. Vol. 18, No. 4, pp. 303-319.Ullman, A. (1985) “Data in Search of a Wallage, P. (2000) “Assurance on Sus- Theory: A Critical Examination tainability Reporting: An Audi- of the Relationships among Social tor’s View”, Auditing: A Journal Performance, Social Disclosure, of Practice and Theory, Vol. 19 and Economic Performance of (Supplement), pp. 53-65. U.S. Firms”, Academy of Manage- Watts, R. (2003) “Conservatism in Ac- ment Review, Vol. 10, No. 3, pp. counting Part I: Explanations and 540-557. Implications”, Accounting Hori-Uzumeri, M. V. (1997) “ISO 9000 and zons, Vol. 17, No. 3, pp. 207-221. Other Metastandards: Principles Wokutch, R.E. & Spencer, B.A. (1987) for Management Practice”, Acad- “Corporate Saints and Sinners: emy of Management Executives, The Effects of Philanthropic and Vol. 11, No. 1, pp. 21-36. Illegal Activity on OrganizationalWaddock, S.A. & Graves, S.B. (1995) Performance”, California Man- “Attraction or Repulsion: How agement Review, Vol. 29, No. 2, Institutional Owners React to Cor- pp. 62-78. porate Social Performance”, Man- Worrell, D.L., Davidson III, W.N., & agement Research News, Vol. 18, Sharma, V.M. (1991) “Layoff No. 2, pp. 20-24. Announcements and Stockholder________ & ______ (1997) “The Cor- Wealth”, Academy of Manage- porate Social Performance- ment Journal, Vol. 34, No. 3, pp. Financial Performance Link”, 662-678. Strategic Management Journal,