corporate environmental crime

Upload: ivonne-duymovich

Post on 03-Jun-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/12/2019 Corporate Environmental Crime

    1/49

    0091-4169/13/10301-0231TH E JO U R N A L O F C R I M I N A L LA W & C R I M I N O LO G Y V o l. 1 0 3, N o . 1Copyright 2013 by Northwestern University School of Law Printed nU S A

    CRIMINOLOGYAN EMPIRICAL ASSESSMEN T OF

    CORPORATE ENVIRONMENTAL CRIMECONTROL STRATEGIES

    SALLY S. SIMPSON,* CAROLE G IBBS,MELISSA ROR IE, * LEE ANN S LO CU M ,

    MARKA.COHEN ,* MICH AE L VANDENBERGH******

    Sally S. Simpson is Professor of Criminology and Criminal Justice at the University ofMa ryland. In addition to research on corporate crime control, her research focuses on white-collar crime data and measurem ent, gender and crime, and criminological theory. She is thecoauthor (with M ichael L. Benson) of White-Collar Crime: An Opportunity P erspective.i Carole Gibbs is an Assistant Professor at Michigan State University with a jointappointment in the School of Criminal Justice and the Department of Fisheries and Wildlife.Her research interests include environmental justice and corporate and white-collar crimesthat harm the environment. Recen t publications app ear in the British Journal ofCriminologyandC riminology and Public Policy.* * Melissa Rorie is a doctoral candidate in the Department of Criminology and CriminalJustice at the University of M aryland, C ollege Park. She is also a Research Assistant at theNationa l Institute of Justice. Her research interests include corporate crime, regulatorypolicy, white-collar crime, and theory testing.Lee Ann Slocum is an Assistant Professor in the Department of Criminology andCriminal Justice at the University of Missou riS t. Louis. Her research interests includeecological differences in the use of formal social control, stability and change in offendingover time, and research methods. Recent work has appeared in Criminology, theJournal ofQuantitative Criminology,and theJournal of Research in Crime and Delinquency.

    Mark A. Cohen is Professor of Management and Law at Vanderbilt University andUniversity Fellow at Resou rces for the Future. He previously served as Chairman of theAm erican Statistical As socia tion's Com mittee on Law and Justice. His research interestsinclude enforcement of environmental regulations, corporate crime and punishment,information disclosure and voluntary overcompliance, and the cost of street crime.

    l i 1 4 1 Michael Vandenbergh is Professor of Law, Director of the Climate ChangeResearch Network, and Co-Director of the Energy, Environment and Land Use Program atVan derbilt University Law School. He has served as Chief of Staff of the Env ironmentalProtection Agency and as a partner in a leading national law firm. His research explores therelationship between formal legal regulation and informal social regulation of individual and

  • 8/12/2019 Corporate Environmental Crime

    2/49

    232 SALLYS. SIMPSON, ETAL. [Vol. 103

    Corpo rate illegality is often attributed to greed by corporate managersand insufficient legal safeguards. Und erlying this argum ent is an explicitcritique of corporate crime regulatory systems. Yet there is little systematicinvestigation of the relative merits of different types or components ofcrime-control strategies; research comp aring more punitive comm and-and-control strategies w ith self-regulatory approach es is particularly lacking.In this Article, w e assess these rimeprevention-and-control mecha nisms inthe context of individual and situational risk factors that m ay increase thelikelihood of illegal behavior in the environmen tal arena. We use datadrawn from two groups of business man agers wh o participated in afactorial survey using vignettes) measuring their intentions to participatein two types of environme ntal offenses. Gen erally, results show that themo st effective regulatory levers are 1) credible legal sanctions and 2) thecertainty and severity of informal discovery by significant others in thefirm. We conclude by discussing the implications of our findings forregulatory policy and strategy, and for efforts to account for the role ofsocial norms in corporate environmental compliance.

    I. INTRODUCTIONAs a subtype of white-collar crime, corporate crime is typicallyunderstood to involve illegal behavior by firms and their agents (executivesand managers) in the pursu it of corporate benefit.' Criminologistsrecognize that even though corporations as juridical persons can be chargedwith illegal activity, corporations per se do not act. Rathe r, managersmake decisions and act on behalf of the company. As corporate acto rs,

    managers also are potentially subject to sanctions for their participation inor know ledge of corporate illegality.^ Enforcement provisions forenvironmental crime allow criminal prosecution, in addition toadministrative and civil sanctions, against both corporations andresponsible corporate officers.^Most corporate crime research focuses on firm, industry, and manager

    ' See JOHN BRAITHWAITE, CORPORATE CR IME IN THE PHARMACEUTICAL INDUSTRY 6(198 4). This definition does not preclu de the notion that self-interest may be an indirectcause and consequence of corporate crime in that managers who problem solvesuccessftjUy, albeit illegally, may reap rewards as an aftereffect.

    See Jennifer Arlen, The Potentially Perverse E ffects of Corporate Criminal Liability,23 J. LEGAL STUD. 833, 834 (1994); Urska Velikonja, Leverage, Sanctions, and Deterrenceof AccountingFraud 44U.C.D VISL . REV . 1281,12 84 2011).^ See Dorothy Thom ton et al.. General Deterrence and Corporate Environmental

  • 8/12/2019 Corporate Environmental Crime

    3/49

    2013] CORPORA TE CRIME CONTROL 233attributes to differentiate offenders from nonoffenders.'' While thisapproachisareasonable one,itoften leavesout animportant characteristicassociated with company (and employee) compliance: the regulatoryenvironment. Specifically, the regulato ry environm ent influencesandshapes criminal opportunities through punishment(or thethreatofit)andsocialization.'

    Putatively, command-and-control policiescompliance mies imposedand policed bythe govemment withanemphasison punitive sanctionsfor violatorsinfluence corporate crime because corporate managersareinstmm ental actors.* Decisions andactions flow from a cost-benefitassessment of both the pecuniary and nonpecuniary pros and consassociated w ith illegal activity. If the benefitsofcrime are high and the riskof discovery and punishmentis low, then criminal opportunities increaseasactors perceive less risk associated with illegal activities.^

    Another regulatory strategy shifts the primary mechanism ofcompliance away from thegovemmentto theorganization itself andtoindividual actors within it. This approach is less reliant on formalregulation (althoughthegovem ment often playsa secondary role through enforced self-regulation) and builds on what Braithwaite hascalleda

    * See MARSHALL B.CLINARD PETER C. YEAGER, C ORPORATE CRIME 43-53 (1980);EDWIN H. SUTHERLAND, WHITE COLLAR CRIME 17-28 (1949); DAVID WEISBURD ET AL.,WHITE-COLLAR CRIME AND CRIMINAL CAREERS143^9 (2001); CindyR.Alexander &MarkA. Cohen, WhyDoCorporations Become Criminals? O wnership, Hidden Actions,andCrimeasan Agency Cost,5J.CORP. F IN.1, 2-5 (1999); Krisfy Holtfreter,IsOccupationalFraud Typical White-Collar Crime?A Comparison of Individual andO rganizationalCharacteristics,33J.CRIM. JUST. 353, 354-56 (2005).

    ^See MICHAEL L. BENSON SALLY S. SIMPSON, WHITE-CO LLAR CRIM E: ANOPPORTUNITY PERSPECTIVE 193-94 (2009); Henry C. Finney Henry R. Lesieur,AContingency Theory of Organizational Crime, in 1 RESEARCH INTHESOCIOLOGYOFORGANIZATIONS255,255 (Samuel B. Bacharach ed., 1982).

    ^Seegenerally NEAL SHOVER ANDY HOCHSTETLER, CHOOSING WHITE-COLLAR CRIME168 (2006) (developing theauthors' theory, which assumes a rational actor model, andopining that sanctions do not work because the comm and-and-control model is notsuccessfully implemented); Gilbert Geis,IsIncarcerationanAppropriate SanctionfortheNonviolent White-Collar Offender? in CONTROVERSIAL ISSUES IN CORRECTIONS 152(Charles B. F ields ed., 1999) (arguing yes).

    ' See Mark A. Cohen Sally S. Simpson, TheOrigins of Corporate Criminality:Rational Individual andOrganizational Actors in DEBATING CORPORATE CRIME 33, 36(WilliamS.Lofquistet al. eds., 1997) [hereinafter Cohen Simpson,Origins of CorporateCriminality](extendingtheeconomic modelto incorporate nonpecuniary costsandbenefitssuch as informal reputation sanctions ); Mark A. Cohen, Environmental CrimeandPunishment: Legal/Economic TheoryandEmpirical EvidenceonEnforcement ofFederalEnvironmental Statutes,82J.CRIM L . CRIMINOLOGY1054,106 3-64 (1992) [hereinafterCohen, Environmental Crime and Punishment] (providinga formal economic modelofth e

  • 8/12/2019 Corporate Environmental Crime

    4/49

    234 SALLYS SIMPSON ETAL. [Vol.103 family model ofcrime control.^

    Good corporate citizensare firms whose managers, when confronted with corporatecriminal opportunities, will be guided by a sense of rightandwrong, by theirunderstandingofhow othersarelikelyto view their behavior, and by theextenttowhich they think the discovery of these acts would bring shame on their companies.

    Effective self-regulation byfirms (ethics program s, intem al compliancemechanisms,andsen sitivityto informal sanctions) should narrow criminalopportunities.

    Inthe corporate crime literature, therehasbeen extensive discussionand debate about different regulatory strategiesbut far too little systematicinvestigation of the relative meritsofeach, and few have taken into accountthe range of solutions that can be included in regula tory policy. ^Consequently, scholars andpolicymakers know very little about whatworks, what doesn't, andwh at's promising regarding corporate crime-control strategies. Inthecurrent study,weoffer some em pirical insigh tinto this question.

    In this Article,weexamine theprevention andcontrol ofcorporateenvironmental crime in the context of individual and firm-levelcharacteristics that have been linked conceptually and empirically tocorporate crim e. Specifically, we focus on the extent to wh ich decisionsbymanagers to violate environm ental laws areaffected by command-and-controlorself-regulation preven tion-and-control strategies, controllingforknown risk factorsforcrime. This research improvesonthe prior literaturein several way s. M uchofthe corporate crime literature relies heav ilyonofficial data sources. As criminologists are well aware, officialobservationsarelimitedtoillegal acts recordedbyenforcement agentsandneglect those acts that do not cometothe attentionofauthorities. Ofequalimportance, these data sources do notallow researchers to learn whatmanagers are actually thinking , leaving the intra-organizationaldecisionmaking process virtually uninvestigated. The current studyaddresses both of these issues byusing data from a factorial surveytoexamine managerial decisionmaking withinacorporate context. Ourgoalistodetermine the extenttowhich regulatory strategies are effective in the

    SeeJOHN BRAITHWAITE, CRIM E, SHAM E AND REINTEGRATION54-68 (1989).' Sally S. Simpson et al..TheSocial Control of Corporate Criminals: ShameandInformal Sanction Threats inOF CRIME CRIMINALITY: THE USE OF THEORY IN EVERYDAY

    LIFE 141,1 42 (Sally S. Simpson ed., 2000).' See NEIL GUNNINGHAM ET AL..SMART REGU LA TION: DESIGNING ENVIRONMENTAL

    POLICY37-88 (1998).'' LAWRENCEW.SHERMAN ET AL.,N A T 'L I N ST .OFJUSTICE, P REVENTING CR IME: WH AT

    WORKS, WHAT DOESN'T, WHAT'S PROMISING(1996),available at https://www.ncjrs.gov/pdf

  • 8/12/2019 Corporate Environmental Crime

    5/49

    2013] CORPORA TE CRIME CONTROL 235context of situational and individual pushes/pulls toward illegal behavior.

    In Part II of this Article, we describe the regulatory context and reviewprevious research on environmental noncompliance. We focus particularlyon organizational and individual factors that increase the risk of crime. InPart III, we describe the current research design and research participants.Partrvcontains our analysis and results. We conclude,inPart V, withadiscussion of the findings, particularly their implications for successfulregulatory regimes.n. PRIOR LITERATURE

    A. REGULATORY STRATEGIES AND CORPORATE OFFENDINGRegulatory strategies often overlap. Regulatory instmmentsandinstitutionsareinterconnected,' and some strategies, suchasresponsiveregulation,arebuilt aroundtheargument that regulatory policy shouldtake neither a solely deterrent nor a solely cooperative approach. '^Although it is somewhat simplistic to classify regulation into distincttypes, * Gunningham, Grabosky, andSinclair argue that it isusefiiltoexamine both the prevention and control capacitiesofdifferent regulatorystrategies given that aparticular instmment which may appear attractive,when lookedat on itsown, may work quite differently when introducedalongside others. '^ Below, weidentify the keycomponents of tworegulatory strategies (command-and-control and self-regulation) andhighlight how eachisexpectedto orhas been showntoaffect corporatecrime prevention and control.'* Inaddition, we discuss the important role

    ' SeeG U N N I N G H A M E T A L . ,supranote 10, at 37-38. Vibeke Lehmann Nielsen & Christine Parker, Testing Responsive Regulation inRegulatory Enforcement 3REG. & GOVERNANCE 376, 376 (2009).'*See IAN AYRES&JOHN BRAITHWAITE,RESPONsrvE REGULATION: TRANSCENDING THE

    D E R E G U L A T I O N DEBATE17-18 (1992).' G U N N I N G H A M E T A L . ,supranote 10, at 132. SeeCohen & Simpson,Origins of Corporate Criminality supranote 7, at 34-35.Theregulation literature is cross-disciplinary andextensive. There aremany other relevantcomponentsof regulatory policy that couldbe considered here, suchas theinfluenceofnongovernmental organizations and corporate gatekeepersonfirm compliance. See e.g.

    JOHN C.COFFEE, JR., GATEKEEPERS: THE PROFESSIONS AND CORPORATE GOVERNANCE 15-54(2006); Bridget M.Hutter &CliveJ.Jones, Business Risk Management Practices:TheInfiuence of State Regulatory Agencies and Non-State Sources17 (Ctr. for Analysis of Riskand Regulationat the London Sch.of Econ. and Political Sci., Discussion Paper No. 41,2006). Other componentsto consider are the barrierstocompliance posedbyregulatoryaccretion. See generally J.B. Ruhl &James Salzman, Mozart and theRedQueen: TheProblem of Regulatory Accretion in theAdministrative State 91 G E O. L.J. 757(2003)

  • 8/12/2019 Corporate Environmental Crime

    6/49

    236 SALLYS SIMPSON ETAL. [Vol.103of informal sanctions either asacontrol m echanism that can be triggered bycommand-and-control interventions or as complementary with self-regulatory strategies.

    1.Command and ControlIn command-and-control strategies, legal authorities dictatethetermsof compliance, relyingon thethreatof formal legal sanctionstoachievecompliance with those term s.'' ' High detection risk coupled with certainand severe punishments should deter most offenders. Em pirically,however, the storyismore complicated than this. Some research supports

    the contention that punitive sanctions affect firm andplant behavior,butfindings overallaremixed. Cohen ,for instance, finds that Coast Guardinspections and monitoring reduce spills at the firm level (a generaldeterrence effect) andthatthefrequency ofinspectionis more importantthan sanction severity.'^ Simpson, Ga mer,andGibbs find little evidencethat sanctions of any type (e.g., insp ections, informal or formalinterventions) associated with Clean Water Act enforcement inhibitreoffending (i.e., specific deterrence ).'^ Plant-level studies moreconsistently show a specific deterrence effect associated withEnvironmental Protection Agency (EPA) monitoring and enforcement, anda recent reviewofthe empirical literature onenforcement, conductedbyGray and Shimshack, finds both specific and general deterrence associatedwith environmental monitoring and enforcement.^specific mechanisms associated with two regulatory strategies affect the way managers thinkand may behave. Inthis regard, our w ork helpstofillanempirical deficit notedbyHutterand Jones:

    W eknow thatthesourcesofregulationandrisk managementarediversifying, as are thetoolsand techniques employed to manage andregulate risks. What we do nothave is muchempirically informed research abouttherangeofsources influencingthebusiness worldand inparticulartheweightingofinfluence exercisedbythem.

    Hutter & Jones,supra at 1. See Clifford Rechtschaffen, Deterrence vs. Cooperationandthe Evolving TheoryofEnvironmental Enforcement 11S.CAL .L .REV. 1181, 1187-90(1998).'* Mark A. Cohen, Empirical Research on the Deterren t Effect of EnvironmentalEnforcement and M onitoring 30ENVTL .L .REP. 10245, 10246 (2000). See SALLYS.SIMPSON E T A L . . W H Y D OCORPORATIONS OBEY ENVIRON MENTAL L AW?

    ASSESSING PUN ITIVE AN D COOPERAT IVE ST RATEGIES OFCORPORATE CRIME CON T ROL 2(2007),available at https://www.ncjrs.gov/pdfFilesl/nij/grants/220693.pdf.Se egenerally Cohen, supra note 18,at 10246 (providing evidence that environmentalmonitoring and enforcement serves both specific and general deterrence functions); WayneB . Gray MaryE.Deily, Compliance and Enforcement:Air Pollution Regulation in the

    U.S. Steel Industry 31 J. E N V T L . E C O N . M G M T . 96 (1996) (discussing how millcompliance with air pollution regulations was associated with substantial reg ulatory

  • 8/12/2019 Corporate Environmental Crime

    7/49

    2013] CORPORATE CRIME CONTROL 237Scenario-based survey research, which largely focuses onenvironmental and other forms of corporate offending (e.g., bribery, salesfraud, price-fixing), shows that current and prospective managers reportreasonably high expectations that corporate crimes will be discovered bylegal authorities and that ensuing sanctions will be costly, particularly whenindividuals (as opposed to the company) are targeted.^' Thus, comm and-and-control strategies based on discovery and punishment should lowercorporate offending. But once again, the relationship is far fromstraightforward. In one study, threats of formal sanc tions are mediatedthrough individual characteristics such as morality^^ and outcome

    expectations.^^ Formal pun ishments are less relevant once informalconsequences are included in the analysis.-^ *2. Self-Regulation

    Self-regulatory approaches (typically offered as a complementarystrategy in conjunction with government-enforced regulation) presume thatprosocial norms and values coupled with effective intemal compliancesystems (e.g., clear accountability, communication of expectations,effective monitoring, and appropriate reprimands when violations occur)will secure compliance.-^^ Bra ithw aite's family model of self-regulationThe Effectiveness of Environmental M onitoring and Enforcement: A R eview of the EmpiricalEvidence, 5 REV. ENVTL. ECON. & POL'Y 3 (2011) (providing a summary of the empiricalliterature on the impact of environmental monitoring and enforcement on plant/facility-levelcompliance); Benot L aplante & Paul Rilstone,Environmental Inspections and Em issions ofthe Pulp and Pa per Industry in Quebec,31 J.ENVTL.EcO N . &M G M T. 19 (1996) (discussinghow both the inspection and threat of inspection have a strong negative impact on plant-levelpollution em issions); Wesley A. M agat & W. Kip Viscusi, Effectiveness of the EPA sRegulatory Enforcement: The Case of Industrial Effluent Standards, 33 J.L. & EcON. 331(1990) (discussing how water pollution inspection and enforcement have a strong effect onpollution and rates of compliance); L ouis W. N adeau, EPA Effectiveness at Reducing theDuration of Plant-Level Noncompliance, 34 J. ENVTL. ECON. & MGMT. 54 (1997)(explaining that the EPA is effective at reducing the length of time plants are out ofcompliance).

    ^ ' See S AL L Y S . S IM P SO N , C O R PO R A TE C R IM E , L A W , A N D S O C IA L C O N T R O L 3 5 ^ 4 2002).^^ See Raymond Patemoster & Sally Simpson, Sanction Threats and Appeals to

    Morality: Testing a Rational Choice Model of Corporate Crime,30L AW&SOC 'Y REV. 549,554 (1996).^ See N . C raig Smith, Sally S. Simpson & C hun-Yao H uang,Why Managers Fail to Do

    the Right Thing: An Empirical Study of Unethical and Illegal Conduct, 17 Bus. ETHICSQ .633, 638-39 (2007).

    ^ See L ori A. Elis & Sally S. Sim pson, Informal Sanction Threats and CorporateCrim e: Additive Versus Mu ltiplicative Models, 32 J. RES. C RIM E & DELINQ. 399, 414-17(1995).

    ^^ See David B. Spence, The Shadow of the Rational Polluter: Rethinking the Role of

  • 8/12/2019 Corporate Environmental Crime

    8/49

    238 SALLYS SIMPSON ETAL. [Vol. 103rejects the economically rational conception of the firm and its managersfound in command-and-control approaches to crim e control.^* Inste ad,company self-regulation also accounts for the notion of organizationalsocial responsibility and the prosocial norms and ethical values of companym anagers. Thus, m anag ers ' perceptions of the ethical climate of a firmshould affect their own offending inten tions.

    Evidence suggests that managers w ho believe that the corporateculture is tolerant of illegality are more likely to violate regulations.^^Similarly, a recent meta-analysis found that ethics codes supported andenforced by top management have a positive, significant effect on ethicaldecisionmaking andconduc t w ithin companies.^^ Resea rch also indicatesthatapositive com pliance cu ltureatthe firm level mayb etransm itted froma paren t compan y to the plant ^'

    3. Informal SanctionsInformal sanctions (e.g., extralegal costs) are regulatory leversassociated with both firm self-regulation and command-and-controlstrategies, depending on the m echanism that inhibits crime . N egativepublicity is a case in point. M ultiple sources of negative publicity canaffect corporate (and manager) actions or outcomes, including:environm ental activism,^ m andatory firm disclosure,^'orformal charges.^^Generally, studies support the notion thatbadenvironm ental n ew s affects afirm 's reputation and m arket perform ance. H owev er, the literatureism ixedas towhen in thelegal process rep utational damage is most salient (noticeof pending charges, case announcement, processing, orcase resolution);^^

    ^* S ee BRAITHWAITE,suprano te 8, a t 13 3 ^ 0 .^^ See Patemoster & Sim pson, supra note 22 , at 556. In addition, man agers with lowerethical standards and m anagers ordered to do so are more likely to violate regu lations. See

    SIMPSON,supranote2 1 ,a t 41-42 .^* N atalie M arie Sc hell-B usey, The De terrent Effects of Ethics Codes for CorporateCrim e: A M eta-A nalysis 85 (2009) (unpub lished Ph.D. dissertation, Universify of M aryland,College Park) (on file w ith Digital R epository at the Universify of M aryland, College Park).^' See Gray & Deify,supranote 20, at 100.^^ See Neil G tmningham et a l . Social License and Environmental Protect ion: WhyBusinesses Go Beyond Com plianc e 29 L A W & Soc . INQUIRY 307, 318, 322 (2004); Rober tA . Kagan e t a l. . Explaining Corporate Environmental Performance: How Does Regulat ionMatter? 37L A W& Soc 'Y REV. 51, 71-72 (2003).^ ' See Shameek K onar & Mark A . Cohen, Inform at ion As Regulat ion: The Ef fect ofCom m uni ty Right to Know Laws o n Toxic Emissions 32 J.ENVTL. ECO N.&MGMT 109 110(1997).

    See W allace N . Davidson III et al.. Stock Ma rket React ions to Anno unced Corpora teIllegalit ies 13 J . B us .ETHICS979, 985 (1994).

  • 8/12/2019 Corporate Environmental Crime

    9/49

    2013] CORPORA TE CRIME CONTROL 39whether stock prices are differentially responsive to civil, criminal, orregulatory moving agents;^'* or even if negative stock reactions are bestunderstood as reputational costs delivered by the market or costsprimarily imposed by the legal community.^^

    When reputational damages stem mainly from formal legalproce edings, this can be seen as part of a deterrence strategy. However,informal sanctions also impose stigmatic, commitment, and attachmentcosts for managers wh o violate the law. * These effects may be a directconsequence of formal san ctio ns or completely unrelated to formalproceedings. In a study of corporate offending intentions. Elis and Simpsonfound inhibitory effects associated with the certainty of intemally imposedinformal sanctions (shame) and extemally imposed informal sanctions (lossof respect from family, friends, and business associates). ^ Importantly, thethreat of both individual and firm reputational damage had an inhibitoryeffect. But these effects were independent of and tended to trump formalsanction risks (which were relatively unimportant sources of deterrence).

    Although the literature is slim and contradictory, there is evidence thatfirm-level stigmatic consequences trickle down to responsible managers.^'In a study of all SEC and DOJ enforcement actions brought betweenJanuary 1978 and September 2006 against 788 firms in which financialmisrepresentation occurred,Karpoff Lee, and M artin report that 93 % of allexecutives and 96% of other employees identified as legally responsible forthe behavior were fired for reasons that are directly related to theirmisconduct. '' Job loss was more likely when misconduct was particularlycostly to shareholders and when offenders faced strong govemance

    ^ * See Bruce Mizrach & Susan Zhang Weerts,Does the Stock Market Punish CorporateMalfeasance? A Case Study of Citigroup, 3 CORP. OWNERSHIP & CONTROL 151, 153-54(2006) (serving as an example of how reputational consequences can flow from differentmoving agents).

    See Jonathan M. Karpoff et a l . The Reputational Penalties for EnvironmentalViolations: Em pirical Evidence,48 J.L. &EcON .653, 665-68 (2005).

    See R aymond P aternoster, The Deterrent Effect of the Perceived Certainty andSeverity of Punishment: A Review of the Evidence and Issues, 4 JuST. Q. 173, 210 (1987)(citing Kirk R . Williams & R ichard Hawkins, Perceptual Research on General Detterence:A Critical Review,20LAW& Soc'YR E V . 545, 568 (1986)). See Kjrk R. Williams & R ichard Hawkins, The M eaning of Arrest for ifeAssault, 27

    CRIMINOLOGY 163,166 (1989).'* See Elis & Simpson,supra note 24, at410-11 .^' See Cindy R. Alexander, On the Nature of the Reputational Penalty for CorporateCrime: Evidence,42 J.L. & EcON . 489 , 523 (1999); Jonathan M . Karpoff et al.. The Cost toFirms of Cooking the Books,43 J.F I N .&QUAN TITATIV E AN ALYSIS581 ,605-07 (2008). Jonathan M. Karpoff et al.. Th e Consequences to Managers for Financial

  • 8/12/2019 Corporate Environmental Crime

    10/49

    240 SALLYS. SIMPSO N, ETAL. [Vol.103stmctures.' From this literature,we expect that command-and-controlaswell as self-regulatory strategies will benefit from accounting for theinformal stigmatic coststothe individual.B . RISK FACTORS FOR CORPORATE OFFENDING

    There are many different emp irically identified risk factors forcorporate crime. These factorsareimportantin ourstudy because effectiveregulation should minimizethelikelihoodofcriminal beha viorin the faceof pressuresandpredilections. B elow,we summarize someof the knownrisk factorsfor corporate environmental crime with the caveat that manyofthese same risk factors also are associated with offending by companiesmore generally.

    Looking firstatfirm characteristics, some research has found financialstrain (measured in different ways)to significantly increasethelikelihoodthat firms, plants, and managers will violate environm ental law s and/orincrease pollution levels.''^ Invigne tte studies specifically, after controllingfor individual-level predictors, managers are significantly more likelytoengagein price-fixing, bribery, fraud,orEPA violationsiftheactwill givethe organization an edge over foreign competitionor the actwill resultinsubstantial savings for the firm. *^ When firm profits are slowing ordeclining, managers and employees may resort to criminal practices toattain perfonnance goals. *^In other studies, however, firm profit either is unrelated toenvironmental (and occupational health and safefy) violations'*^ or has apositive effect onoffending. * Simpson ,for example, finds that managers'intentionstooffend were higher whenthefirm w as depictedasgrowingitssales.''^ Thu s, offending may be related to both financial declineand

    ' Id at194.* SeeCLINARD YEAGER, supra note4, at 128-29; CindyR. Alexander MarkA.Cohen,NewEvidence on the Origins of Corporate Crime, 17MANAGERIAL DECISION

    ECON. 421 ,421 (1996); Kaganet al.,supra note 30, at 51-90; Marie McKendallet al.Corporate Governance and Corporate Illegality: The Effects of Board Structure onEnvironmental Violations,7INT'LJ .ORG. ANALYSIS2 01 , 203 (1999).*'See Patemoster & Simpson,supranote 22,at557-59. See generallyNeal Shover & Kevin M. Bryant, Theoretical Explanations of Corporate

    Crime inUNDERSTANDING CORPORATE CRIMINALITY 141, 154 (Michael B. Blankenship ed1993).See CharlesW. L.Hilletal..AnEmpirical Examination of the C ausesofCorporateWron gdoing in the Un ited States,45 HUM. REL. 1055, 1070-71 (1992).

    ' SeeMarieA.M cKendall John A.Wagner,III,Mo tive, Oppo rtunity, Choice, andCorpora te Illegality,8ORG. SCI.624, 625-26, 638 (1997).

  • 8/12/2019 Corporate Environmental Crime

    11/49

    2013] CORPORATE CRIME CONTROL 241growth. Corporate observers suggest that a lack ofpredictabilitymayunderlie bothof these situations, especiallyifthe changeisunexpectedorrapidinnature.'*^

    The literature also predicts a link between intra-organizationalstructures and offending. Within com panies, decisions follow particularlines of communication andresponsibility. M anagers have shown atendencytoobey authority, even when orderedtobehave unethicallyorviolatethelaw. *' Thisistrue especiallyformiddle-level managers w hoareresponsible for carrying out orders but who have relatively littledecisionmaking authority vis--vistopm anagement.^Although organizational characteristics are often associated withcorporate crime because they provide opportunity, context,orm otivationsfor offending,asmentioned previously, managersnot companiesm akedecisions. Managerial decisions mightbe affected byindividual norms^'and differences in traits such as impulsivity, hubris, desire forcontrol,Machiavellianism,andself-serving bias.^^ Although ev idenceonthelinkbetween corporate crimeand lowself-control isweakatbest, * empiricalfindings support anassociation between some ofthese other individualtraits and illegality and/or other negative business outcomes.'^

    * MICHAEL E. PORT ER, COMPETITIVE STRATEGY: TECHNIOUES FOR ANALYZINGINDUSTRIES AND COMPETITORS324 (1980).

    ' HERBERTC. KELMAN V.LEEHAMILTON, CRIMESOFOBEDIENCE: TOWARDASOCIALPSYCHOLOGY OF AUTHO RITY AND RESPONSIBILITY(1989).

    ' See MARSHALL B.CLINARD, CORPORATE ETHICSANDCRIME: THERO L E OF MIDDLEMANAGEMENT21-23 (1983).

    See MichaelP.Vandenbergh,Beyond Elegance:A Testable Typology of Social Normsin Corporate Environmental Comp liance 22STAN. ENVTL.L.J. 55, 63-67 (2003). See Paul Babiaketal..Corporate Psychopathy: Talking theWalk 28BEHAV. SCI.&L.174, 190 (2010).

    See MICHAEL R. GOTTFREDSON TRAVIS HIRSCHI, A GENERAL THEORY OFCRIME180-201 (1990) (discussing white-collar crime); Travis Hirschi Michael Gottfi-edson,TheSignificance ofWhite-Collar Crime for a General TheoryofCrime 27CRIMINOLOGY359,360-62(1989) .

    ' ' 'See WEISBURD ETAL., supra note4, at 187-88; Michael L.Benson ElizabethMoore,AreWhite-CollarandComm on Offenders the Same?:AnEmpiricalandTheoreticalCritique of a Recently Proposed G eneral Theory of Crime 29 J.RES. CRIME&DELINQ .251 ,260-63 (1992); Sally S. Simpson Nicole Leeper Piquero, Low Self-ControlOrganizationalTheory andCorporate C rime 36LAW & Soc'YREV. 509, 531-33 (2002).

    See KatherineA.DeCelles MichaelD.Pfarrer, HeroesorVillains? Corruptionandthe Charismatic Leader 11 J. LEADERSHIP ORGANIZATIONAL STUD.67,69-70 (2004);MatthewL. A.Hayward DonaldC.Hambrick,Explaining thePremiums Paid for LargeAcquisitions: Evidence of CEO Hubris 42ADMIN. SCI.Q. 103, 106-10 (1997); NicoleLeeper Piquero,M. LynExum SallyS.Simpson, Integrating theDesire-for-Control and

    aCorporate Crime Con text 22 Q .

  • 8/12/2019 Corporate Environmental Crime

    12/49

    242 SALLYS SIMPSON ETAL. [Vol. 103In sum,thebodyofevidence rega rding the specific leversofcorporate

    crime control is limited and inconclusive. Im portantly, little is knownabout how effective different regulatory strategiesorcomponents are in thecontext of known or suspected risk factors^both organizational andindividual. To address this issue, w erely on data drawn from a factorialsurvey administered to two samples of corporate employees. The firstsam ple includes corporate ma nagers recm ited aspartofa National Instituteof Justice (NlJ)-funded study on corporate environmental noncompliance.The second sample includes employees ofpublic and private companies,drawn from a broader sample ofthe popu lation. Both sets ofp articipantsresponded to web-based surveys that depicted tw o hypothetical scenariosdesignedtoassess individ uals' p ropensitytoengagein significant pollutionviolations (e.g., discharging toxins into waterways) and less seriousenvironm ental offenses (e.g., ignoring anEPA compliance order).

    I I I . M E T H O D SFactorial surveys combine experimentally manipulated hypotheticalscenarios (vignettes) with survey questions to measure respondents 'intentions, decisions, attitudes, or judgm ents.^* These desig ns, unlike m oretraditional survey techniques, allow researchers tomanipulate a full rangeof circiunstances that may affect adecision essentially taking into account the complexity and richness in the way people approach decisions andevaluat ions. Thedesign also avoids some of the temporal orderinga ndperceptual instability problems associated with other research designs.'^Vignette surveys can be less threatening methods of data collection thanself-report surveys when th e subject matter is sensitive, such as whenrespondents are queried about unethical, criminal, or deviant behavior.

    Although vignettes have been used extensively in thesocial scienc es^'an d^ ' See PETER HEN RY ROSSI & STEVEN L . N OCK, MEA SURING SOCIAL JUDGMENTS 9 -13(1982). Edw ard D. Weber et a l . The PC Vignette Generating Program (1988) (on file with theUniversify of Massachusetts Social and Demographic Research Institute, Amherst, MA).^* See Linda Saltzman et al.. Deterrent and Experiential Effects: The Problem of CausalOrder in Perceptual Deterrence Research 19 J.RES. CRIME&DELINO. 172,174 (1982).

    In crim inology, vignette studies w ere used to evaluate the appropriateness of corp oratecrime punishments, .see Joan L. Miller et al.. Felony Punishments: A Factorial Survey ofPerceived Justice in Criminal Sentencing 82 J. CRIM. L . & CRIMINOLOGY 396, 396-415(1991); public perceptions of white-collar crime seriousness, see James Frank e t a l .Sanctioning Corporate Crime: How Do Business Executives and the Public Compare? 13A M. J . CRIM. JUST. 139, 139^1 (1989); and offending intentions , see Steven Klepper &Daniel Nagin , The Deterrent Effect of Perceived C ertainty and Severity of Punishm entRevisited 27 CRIMINOLOGY 72 1, 729 (19 89) ; George Loe w ens tein et a l . The Effect of

  • 8/12/2019 Corporate Environmental Crime

    13/49

    2013] CORPORATE CRIME CONTROL 243business to collect useful information about topics such as surveyparticipation* and consumer preferences,*' they also have some drawbacks.Critics have raised conce ms about: the link between reported intentions andactual behavior and whether the relationship is sensitive to the respon den t'ssex and the situation analyzed;*^ the extent of social desirability bias inresponses, especially in the constant-variable value vignettes where allrespondents read identical scenarios;*^ whether there are ord er effects;*and whether scenario-based research is valid and reliable;*' among otherissues. Poorly considered dimensions and components of the scenarioscontribute to validity prob lems.

    In this study, to increase data validity, we draw extensively from theempirical and theoretical literatures to identify relevant attributes and levelsfor the vignettes. W e also asked environmental professionals and people inbusiness to review drafts of the vignettes and the instrument was revisedaccordingly.** We experimentally rotated vignette items within thescenario dimensions to lessen social desirabilify bias . To minim ize thebiasing effect of vignette order, we randomly allocated items to respondentsand asked them to imagine themselves in the position of the vignetteactor.*'' Order effects are more likely when respondents have little

    445 (1997); Patemoster & Simpson, supranote 22, at 558; N. Craig Smith et al, supranote23, at 645.

    ' See Robert M. Groves et al., A Laboratory Approach to Measuring the Effects onSurvey Participation of Interview Leng th Incentives Differential Incentives and Refusa lConversion 15 J. OFFICIALSTAT.251 (1999). See Alice Gr0nh0j & Tino Bech-Larsen, Using Vignettes to Study F amily

    Consumption Processes 27PSYCHOL.&MARK E TING445 (2010). See Stefanie E ifier, Evaluating the Validity of Self-Reported Deviant Behavior UsingVignette Analyses 41 QUALITY &QUANTITY 303, 306-10 (2007); M. Lyn E xum et al. SelfReported Intentions to Offend All Talk and No Action? 37 AM. J. CRIM. JUST. 523, 534(2011).'' Seegenerally Gerald. F. Cavanaugh & David J. Fritzsche, Using V ignettes in BusinessEthics Research in 1 RESEARCH IN CORPORATE SOCIAL PERFORMANCEAND POLICY 27 9- 93(Lee E . Preston ed., 1985); Maria F. Femandes & Donna M. Randall,The Nature of SocialDesirability Response Effects in Ethics Research 2Bus.ETHI SQ . 183 (1992).''* See K atrin A uspurg & A nnette Jackie,First Equals M ost Important?: Order Effects in

    Vignette-Based Measurem ent 19-20 (Inst. for Soc. and E con. Research, Working Paper No.2012-01).' ' See Stefanie E ifler, Validity of a Factorial Survey Approach to the Analysis ofCriminal Behavior 6 METHODOLOGY 139, 140 (201 0); Jam es W eber,Scenarios in BusinessEthics Research: Review Critical Assessment and Recom mendations 2 Bus. ETHICSQ .137,145-46(1992).^ See K elly D. Wason et al.. Designing Vignette Studies in Marketing 10

    AU STRALASIAN MARK E TING J. 41,53 2002).

  • 8/12/2019 Corporate Environmental Crime

    14/49

    244 SALLYS. SIMPSON ETAL. , [Vol. 103knowledge or care little about the topic.^ Our study explicitly sampledenvironmental experts (discussed below), and we asked respondents toreporton their experiences withand attitudes about depicted behaviorsaspartofthe survey. Although the juryisstill out, research alsohasshownareasonable correlation between reporied intentions andbehavior.^^ Giventhe sensitive subject matter (corporate offending) aswellas ourattentiontomethodological concems in the design of the instrument, we believethefactorial survey methodis areasonableandvalid approachto ourresearchquestion.A. SCENARIO CONSTRUCTIONOneof the first steps in factorial survey constructionis to determinethe domain ofthe judgmentordecision. Thevignette domain consistsofdimensions believedtoaffect thema nager's decisiontoengageincorporateoffending. Guidedby amodified rational choice theory^ and the corporatecrime empirical literature,anumber of pushes and pulls toward crimeat theindividual and corporate level are incorporated into the vignette design.Scenarios are created fi-om rotated elements or levels within dimensions.To illustrate,the firm's environmental constraints(onedimension) providecontextual information to the respondent aboutthe economic environmentin which thecompany is conducting itsbusiness. Thetype of constraint(e.g.,theindustry is losing ground to foreign competitors, the industryiseconomically healthy, the industry is economically deteriorating) israndomly assignedtoeach scenario. Further, management levelis anintra-organizational dimension empirically linked to offending decisions.Pressures on middle managers to achieve corporate goalsoften withtheimplicit message by anymeans necessary ^'and unrealistic performancemetrics^^ create a greater likelihood of corporate crime by midlevelmanagers compared with top management (who generally set companygoals andstrategies for others to achieve). Research has also found thatmanagers adhereto authority structures within firms. Theprobabilityof

    ** Auspurg&Jackie,supranote 64,at 1. See Greg Pogarsky, Projected Offending nd Contemporaneous Rule-Violation:Implications for Heterotypic Continuity 42CRIM INOLOGY 111,115 (2004); Harry Telser&Peter Zweifel, Validity of Discrete-Choice Experim ents Evidence for Hea lth Risk Reduction

    39 APPLIED ECON.69, 72-75 (2007). See Patemoster & Simpson,supranote 22,at553-57.^' CLINARD,supranote 50,at22-23.'^ John Braithwaite, White-Collar Crime Com petition nd Capitalism: CommentonColeman 94 AM .J. Soc . 627, 629 (1988).'^ Joseph Sanders & V. L ee Hamilton,Distributing Responsibility for Wron gdoing Inside

  • 8/12/2019 Corporate Environmental Crime

    15/49

    2013] CORPOR TE CRIME CONTROL 245corporate crime is higher for those managers affirmatively instmcted tobreak a given m le. Thus, we varied the locus of control in the scenariosbased on whether or not the employee was ordered by a superior to committhe offense. Finally, the ethical tone and culture of a com pany can affecthow managers perceive corporate crime (as acceptable or not) which, intum, may increase or decrease the anticipated rewards/costs of offending.Dimensions that rotate levels of managerial ethics, firm volunteerism, andintemal compliance systems also are randomized w ithin the scenarios. Thespecific dimensions of interest (and randomized levels within each) for thisArticle are listed in Appendix I.

    The survey instrument contains two offending vignettes. Onenoncompliance scenario describes a technical violation (e.g., failure toact/comply with an environmental agency's order) with no indication ofwhether it will affect po llution leve ls. The other depicts a more substantialpollution event (the intentional release of a toxic substance into a localwaterway) that exceeds permitted levels by 200% . Sample scenarios can befound in Ap pend ix II. The vigne ttes are followed by a series of questionsthat relate to a specific scenario, general questions that measurerespondents' opinions and beliefs, and requests for demographic/workinformation about the respondent and his/her current em ployer.B . SURVEY ADMINISTRATION

    As noted previously, the factorial survey was first pretested,redesigned, and then vetted with environmental scholars, regulators, andexecutives. The instmm ent was modified to address any remainingconcems and adopted to be administered using a web-based Intemet site.Data collection occurred in two w aves. The survey first was administeredin companies that were part of a larger NlJ-funded grant. * The originalresearch assessed pattems of environmental offending and companyresponses to govemm ental interventions (regulatory, civil, or criminal). Anadditional goal was to look inside the black box of the corporation bystudying managers' perceptions of corporate environmental crime and leamabou t their decisionm aking processes. All firms (whether participants inthe vignette survey or not) were drawn fi-om a sample of U.S.-ownedcompanies in three basic manufacturing industries (steel, pulp and paper,and oil refining). These firms owned manufacturing sites that weredes igna ted by the EPA as major facilities.^^ Of the forfy-eight firms853(1997).

    ' See SIMPSONETAL.,supranote 19.'^ W hether a facility is deem ed ma jor or not is determined by the facility's volume and

  • 8/12/2019 Corporate Environmental Crime

    16/49

    246 SALLYS SIMPSON ETAL. [Vol. 103contacted toparticipate in this study, only three companies agreed to sendthe surveyout to their employees. Due to technical difficulties associatedwith survey administration, one firm asked to withdraw from the projectafter data had been collected. Thus, respondents for the study areconcentratedin twoparticipating companiesonein thesteel industryandthe otherinpulpandpaper.

    The lowrateof participation raises concemsofbiasin thedata. Forexample, if ethical firms were more likely to participate, our resultsmayunderestimate the likelihood of offending and the relationship between itandkeyindependent variables. Toassess potential bias,wecompared eachparticipating company's average size (number of facilities owned, numberof employees), market value (total stockholders' equity), compliance (totalviolations, violation ratepernumberof facilities), and enforcement record(total sanctions, total inspections, inspection ratepernumberoffacilities)tothe seventy-one nonparticipants.^* Thevaluesforeach company were takenby averaging six years of data (1995-2000). We do not providesignificance testsdue to thesmall sam ple sizeforparticipants N =2). Oneparticipating company is much largerand has a higher market value thanthe average company in the sample. It also has a better record ofcompliance, includinga lower violation rate .43 standard deviations belowthe mean for nonparticipating companies). The second participant is alsosomewhat larger thantheaverage com pany,but has arecordofcompliancevery similartothatofnonparticipating companies .01 standard deviationsbelow the mean for nonparticipating com panies). Such firm-levelvariability gives us confidence that responding managers come fromcorporate environments with different environmental recordsone betterthanand onecomparableto the typica l nonparticipating firmin the NUsample.

    Fifty-four respondents from one company and sixteen participantsfrom another reported on their willingness to engage in thenoncompliantbehaviors described in both scenario types, producing 140scenarios. Anadditional fourteen participants reported their behavioral intentions foronescenario. This produced 154 cases for potential analysis. After listwisedeletionofindependent variables,126cases from seventy respondents wereincludedin thefinal analysis.^'Yeager,Industrial aterPollution 18 RIME&JUST.97,122n.37(1993).

    '*SeeinfraAppendixIII,Table1.^ For both samples, only seventeen people who received both scenarios failed torespondtoboth (fourteen in the NUsam ple; threein theTM one sample). Because elevenofthese are missing additional data, these respondents ultimately are excluded from furtheranalysis. Not surprisingly, after dropping thosewho completed onlyonescenario fromthe

  • 8/12/2019 Corporate Environmental Crime

    17/49

    2013] CORPORATE CRIME CONTROL 247Recognizing that respondents in the NIJ study were drawn from alimited sample of large manufacturers, our goal in the second study was totarget a broader set of business managers and gain a larger pool ofrespondents than in the first study. ^ We also preferred poten tialrespondents to have some knowledge about environmental statutes andcompliance requirements. Accordingly we obtained a list of 7,292environmental decisionmakers w ithin a wide variety of o rganizations in theUnited States from TMone, a company that provides targeted databases fordirect mail campaigns. The sampling frame list contained the individual'sname as well as information about the entity for which he worked,

    including the organization's name, address, telephone number, andwebpage (if applicable).From December 2008 to March 2009, Vanderbilt University sentletters to potential respondents on the list indicating their selection forsurvey participation. The letters also provided a link to the web-basedfactorial survey. Of the 7,292 sent, 1,373 letters were retum ed asundeliverable, leaving us with a potential sample pool of 5,919. Toincrease response rates, Vanderbilt researchers sent out follow-up postcardsabout three weeks after the initial letter was sent (from January 2009 to

    April 2009). Seven hund red seventeen individuals logged into the surveysite,for a response rate of about 12%.^' This response rate is not atypical ofthat seen in previous studies on web-based surveys.^ Low response ratesdifferences.

    ^ There were a few m inor changes between the two survey instmm ents. Whenapplicable, differences between items are noted in the text.^' We received about thirfy contacts from individuals about the surveys, includingreasons for participating or not participating in the survey. Mo st of the contacts whoindicated they would not be responding to the survey mentioned a lack of technicalproficiency or not owning a computer. Many who reported not owning a computer werealso retired. Some m entioned that they simply were not interested or that they did not feellike they were appropriate respondents because of their current jobs or work experiences(e.g., did not see them selves as environmental ex perts). We provided technical assistance toindividuals who wanted to respond to the survey but had trouble accessing the website, andencouraged those who felt they were inappropriately contacted to respond with theunderstanding that we would consider their job description and experience when interpretingresults.* See generallyStephen R. Porter & Michael E. Whitcomb, The Impact of Contact Type

    on Web Survey Response Rates 67 PUB . OPINION Q. 579, 583-84 (2003) (comparingdifferent types of contacts and showing that response rates are relatively low regardless ofdelivery condition); Ashok Ranchhod & Fan Zhou, Comparing Respondents of E-mail andMail Surveys: Understanding the Implications of Technology 19 M RKETING INTELLIGENCE& PLANNING 254 (2001 ). Scholars have long noted that web-ba sed surveys have certainfeatures that reduce the likelihood of respo nse. For instance, researchers cann ot includetangible incentives that can increase participation (e.g., pens, stickers), the formatting of websurveys may make the questionnaire appear longer and less professional, respondents may

  • 8/12/2019 Corporate Environmental Crime

    18/49

    248 SALLYS SIMPSON ETAL. [Vol. 103do not necessarily equate tononresponse bias. Ifrespondent characteristicsare similar to nonrespondent characteristics, then the survey responses canreasonably be attributed to the larger target sample.* To assessnonresponse bias, we took a random sample of 500 individuals from the5,919 who received the invitation to participate. For each of these 500people, we gathered additional information regarding the type oforganization (publiclyor privately owned corporation, govemment agency,NGO or civic association, law firm, private consulting firm, or other)andthe fype of profession (environmentally related or not , the size of theentify,and thegenderof the individual. Asshown inAppendix III (Table2 ), we compared respondents tononrespondents on these four dimensions,and found thatthe only significant difference between the twogroups wasorganization size. Respondents came from slightly larger organizations(mean= 14.04employees) than nonparticipants (mean = 9.59employees).^^However,thesimilarities on theother dimensions suggest nonresponse maynotbe amajor problem . Even so, we are cautious with our interpretationand extrapolationoffindings.C. RESPONDENT INFORMATION

    Merging the two sets of respondents yields a total of 237 scenariosfrom 161 individuals.^^

    feel that data integrityis notsecure,and technical issues may affect responsivify. Seeid.at254-56; LindaJ. Sax et al..Assessing R esponse R ates andNonresponse Bias in Web andPaper Su rveys,44RES. HIGHER EDUC.409(2003).^ Sax etal.,Mpra note 80, at409 -32 .

    ^ The rainge on this variable is 1 to 370 emp loyees. Although statistically significant,we question whether thisis ameaningful difference. Both mea ns are on thesmall side and ifwe round9.59 up to 10, both responders and nonparticipants fall within the same businesssize classification accordingto theU .S. Census. Statistics About Business Size, U .S . CENSUSBUREAU (Aug.22,2 012 ), http://www.census.gov/econ/smallbus.html.^ For comparison purposes,thedemographic characteristics of each sample are shownin Table 1,infra. As these descriptions demonstrate,ourstrategy to get abroader range of

  • 8/12/2019 Corporate Environmental Crime

    19/49

    2013] CORPORATE CRIME CONTROLTable 1

    Dem ographics and Perceptions of Compan yEnvironm ental Strategies by Sam ple

    249

    Marital StatusGenderAgeEducationHS degree orequivalent___Sorriejcollege

    yearcollege degreeSome graduate studyGraduate degree

    NU Sample(N 70)94% married93%maleMean = 45.99

    2.74%5.48%57.53%13.70%20.55%

    TMone Sample(N 91)80% married79% maleMean = 54.47

    57.29%30.21%5.21%1.04%6.25%

    Involvement with environmental decisionmakingNot involvedSomewhat involvedRoutinely involvedManagement experienceYears working forcurrent employerYears of businessexperienceManagerial position

    4.29%28.57%67.14%Mean = 18.78Mean = 23.34

    Lower: 17.14%Middle:75.71%Upper: 7.14%

    4.17%10.42%85.42%Mean =13.87Mean = 30.84

    Employee: 5.15%Management: 17.53%Executive: 77.32%Environmental commitment ofrespondent'sflrmExcessiveAbout rightCould use work Poor

    10.00%88.57%1.43%0.00%1Compliance systems in respondent s companyCode of ethicsMandatory ethicstraining1 Random audits1 Anonymous hotline1 Corporate environ.1 mgmt. system ori company policy^'

    i Top mgmt. treats ethicsi and violations seriously

    100%91%26%99 %99%

    91%

    U 9.28%88.66%2.06%0.00%63%20%4%10%38%

    60%

    Total Sample(N=161)85% married84% maleMean = 50.4935.56%20.00%26.67% '6.11%11.67%4.38%18.13%77.50%

    Mean =16.12Mean = 27.39

    9.54%88.60%1.86%0.00%80%51 %14%49%

    65%

    *'' The two samples received different response choices to the question, "What is yourmanagement level?"^ The two samples received different questions about com pany policy. The NU samplewas asked, "Does your current employer have [a] Corporate Environmental ManagementSystem ?" The TMone sample was asked whether their company had a "[cjompany policy

  • 8/12/2019 Corporate Environmental Crime

    20/49

    250 SALLYS.SIMPSON ETAL [Vol. 103

    As reportedinTable 1, theaverage respondentis about fifty yearsofage. Mostaremarried (85 ) and male (84%),andasubstantia l majorifyattended collegeorgraduate school (64%). Theyareexperiencedandloyalworkers, withanaverage career length oftwenfy-seven years, sixteenofwhich have been with their current employer. Abou t 78 report beingroutinely involved with environmental decisionmaking in their respectivecompanies. Most respondents have a positive assessment oftheir firms'ethical com mitm ents. Sixfy-five percentofthe respondents report thattopmanagement treats ethicsand violations seriously,and89% report thattheenvironmental commitment of their firm isjust right. There is morevariationin thekindsofethics trainingand compliance systems utilizedbytheir firms. About 80 ofrespondents report thepresence of acodeofethicsintheir w orkplace, 5 1 % report mandatory ethics training, 14% reportrandom audits, and 49 % reportananonymous hotline.^*D . MEASURES

    1. ependent VariableAfter reading each scenario, respondents were askedtoratehowlikelythey wereto actlikethemanagerin thescenario(0 = nochanceat all to 10= 100 chance),whoinboth scenarios always engages innoncompliantbehavior. Therefore thedependent variable measures therespondent'swillingness or intention to actillegally. Intentions are notdistributedequally across vignette types; individuals were much less likelyto offendwhen giventhe significant noncom pliance scenario than when they weregiven the technical nonco mpliance scenario. Specifically, in the 113

    EPA-order-defiance scenarios, 57.5%(N = 65) ofthe respondents reportedno chanceofoffending. Incomparison, 102 out ofthe 124(82.3%)oftherespondentstothe toxic-release scenario reported nochanceof offending.

    * This seem s fairly typ ical formajor corporations. Arecent LRN ethics study,forinstance, found eight outoften employees reporting that their organization hasa writtencodeofconductorethics. Moreover,a large percent (83%) alsosaythat their managem ent genuinely wantstoprom ote integrity andethicsinthe organization. LRNC O RP.,THEIMPACT OF C O D E S OF C O N D U C T ON C O RPO RA TE C U L TU RE 3 (2006), available athttp://virww.ethics.org/files/u5/LRN ImpactofC odesofC onduct.pdf Thecode of ethicsorcodeofconductis the most common partof the compliance infrastmcture andhotlinesarealso fairly com mo n. Highly regulated industries tend tohave m ore detailed compliancestmctures. D O N N A BO E H M E , F RO M E N RO N TO M A D O F F : WHY M O S T C O R PO R A TECO M P L I A N CEANDETHIC S PRO GRAMSARE PO SITIONED FORFAILURE 28 (2009), available athttp://compliancestrategists.net/sitebuildercontent/sitebuilderfiles/Randl.pdf Weexpectedand foundthe NUrespondents (steel; pulpandpaper)to report more extensive systems than

  • 8/12/2019 Corporate Environmental Crime

    21/49

    2013] CORPOR TE CRIME CONTROL 251Combined, 167 out of the 237 (70.5%) scenarios in the total survey had a no chance of offending outcome reported.^ Given the obvious positiveskew on offending overall as well as by scenario type, we recoded thedependent variable to a binary outcome (0 = unwilling to act like themanager in the scenario; 1 = 10% or more chance of acting like themanager in the scenario and engaging in illegal behavior).^^

    2.Independent VariablesThe independent variables can be sorted into three main categories:relevant control variables, individual/corporate risk factors for offending,

    and variables that capture some aspect of regulation (comm and-and-control,firm self-regulation). Variables are drawn from the vignettes themse lves,^'questions that follow each survey, and demographic information reportedby respondents. Each class of variables is described briefly below.i. Controls

    Although there are a number of potential control variables to includein this analysis, we focused primarily on variables that had a significantbivariate relationsh ip w ith the dependent variable ^ or affected offendingintentions in preliminary analyses, once other variables were included in themod els. Questions ask respondents to assess scenario realism (0 = notrealistic, = realistic) and to rate the desirability ofthebehavior (0 = not atall desirable to 10 = very desirable) and how likely the manager's actionswere to endanger human life and wildlife (0 = no chance at all to 10 =100% chance). We also included measures that controlled for: (1) therespondent's years of business experience; (2) whether the respondent hadpersonally experienced or knew about any of the offense types presented;

    *' O nly people w ith no missing data are included in our analysis. Conse quently, we donot lose any cases across models.^* Altemative coding strategies for the dependent variable (0-10) and analyses (OLS andTobit) were conducted to assess whether the binary coding scheme is defensible. Results aresubstantively the same as those reported with only minor observed differences (resultsavailable on request). Because our primary interest in this study is to determine whe thercertain interventions minimize the likelihood of offending, it made sense for us to truncateall positive responses into a binary 0/1 coding scheme.^' Vignette characteristics entered into the models are dummy variables scored as if thecharacteristic was present in the scenario and 0 if it was not included. The choice of thereference category is determined by theory and ease of interpretation.' SeeinfraTable 2. The two samples received slightly different versions of this question the NU samplewas asked whether they had personal experience with or knew about any of the scenariospresented in the same question while the TMone sample was only asked about personal

  • 8/12/2019 Corporate Environmental Crime

    22/49

    252 SALLYS SIMPSON, ETAL. [Vol. 103(3) scenario type (technical noncompliance versus release of toxins); and(4) whether the respondent was part of the NIJ sample (coded as 1) or theTM one sample (coded as 0).ii . Risk Factors

    Individual-level factorsThe corporate crime literature has identified several individual-levelcharacteristics that might confound the relationship between prevention-and-control strategies and offending intentions. Our respondents wereasked several questions that tap into some of these attributes. Personalself-

    interest was measured by two questions about perceived career benefitsfrom the crime (cmde measures of Machiavellianism) and the level ofexcitement associated with illegal behavior (respondents rated both on 11-point scales where higher scores indicated greater perceived benefits).'^We expect that respondents who perceive more career benefits and greaterthrills will report higher offending intentions. The risk of crime also shouldbe greater when managers do not believe in the common value systemrepresented by the law. Social control theory, for instance, asserts there isvariationin the extent to which people believe they should obey the mies ofsociety . . . . [T]he less a person believes he should obey the m ies, the morelikely he is to violate them. '^ To examine this, we asked respondents theirdegree of agreement with the following statement: Indiv iduals shouldcomply with the law so long as it does not go against what s/he thinks isright. Agreement with such a statement suggests that the respondent doesnot share conventional attitudes about the moral authority of law and istherefore at greater risk for offending.Companv-level factorsRisks associated with the company are captured in two types ofmeasures: economic constraints on the firm and managerial position and

    presented in the same question while the TMone sample was only asked about personalexperience with each scenario. Further, we could not distinguish which of the specificvignettes the NIJ sample had personal exp erience with or knew about. Thu s, for bothsamples, this variable reflects personal experience or knowledge about any of theenvironmental situations presented, including overcompliance scenarios (not discussed inthis pape r). Although the NIJ sample is more likely to report having personalexperience/hearing about these behaviors (mean = 0.896) than the TMone sample (mean =0.577), this result may be due to the more inclusive nature of the NIJ question than a tmedifference in personal experience.'^ The distribution of the perceived thrills variable is skewed to the right, with 197 outof 237 responses at 0 (out of a possible 10). The mean of this variable is 0.43. W eexamined all six models using a dichotomous version of the variable, but results weresubstantively the same. We therefore report the results using the full 10-point scale.

  • 8/12/2019 Corporate Environmental Crime

    23/49

    2013] CORPOR TE CRIME CONTROL 253authorify. In the vignettes, respondents assessed scenario conditionswherein the firm was depicted as economically deteriorating, economicallyhealthy , or losing ground to foreign com petition. In addition, research hasshown that a manager's location within the company affects offending riskas pressures are often placed on middle managers to meet performancegoals regardless of whether these goals can be achieved within theconstraints of the law. Similarly, offending risks are also tied to whethermanagers are in a position to com pel others to act illegally. So managerlevel (middle and upper) and decisionmaking authority (asked bysupervisor to act versus made an independent decision) are included ascompany risk factors; both variables are drawn from the hypotheticalscenarios.iii. Regulation

    Firm Self-RegulationSeveral vignette dimensions are important indicators of companyselfregulation . For exam ple, scenarios contained information about whetherthe depicted company participated (or not) in voluntary EPA pollution-reduction programs; the kind of intemal compliance system at the firm (arange from mandatory training through mandatory self-reporting ofreleases); whether ethics typically guide decisionmaking in the firm; andthe consequences for managers who were discovered by the company to bebehaving in similar activities (graduated consequences from noconsequence at all to the employee w as fired).Informal Sanction RiskWe also created a standardized scale that takes into account theperceived certainfy and severify of three business-related informal sanctions

    directed at the individual. Variables include the perceived likelihood andcost of losing the respect of business associates, loss ofjob, and future harmto jo b prospects if the behavior was discovered informally. This scale alsoincorporates a measure of responden ts' perceptions regarding the likelihoodthat their actions would be discovered by the firm but not by legalauthorities.^Command and Control

    ''' Prior to standardization, scores can range fi'om 0 to 3000. After cen tering, the range isfi-om 1.889 to 1.699. Preliminary analysis revealed a significant difference betw eensamples. TM one respondents perceive lower risks (mean = 1321.113) than do NUresponde nts (mean = 1812.02). How ever, the biserial correlations betwee n the informalsanctions scale and the outcome are similar for the two samp les. The biserial correlationbetween informal business sanctions and offending decisions are NU rpb = -0.38, p < .01;TMone rpb = -0.36, p

  • 8/12/2019 Corporate Environmental Crime

    24/49

    254 SALLYS.SIMPSON, ETAL. [Vol.103After each scenario, respondents were asked toevaluate th eadequacy

    (0= toostrict to 10 = toolenient) ofthe laws goveming thebehav iordescribed in thescenario.'^ In addition, five questions focused on therespondents ' perceptions ofth e formal costsofoffending for theindividualactor. The responses to these five q uestions w ere com bined to formaformal sanction scale.^^ Thescale takes into ac count both resp ond ents'perceptions ofth e likelihood that a given negative outcome (e.g., beingarrested, sued,orinvestigated byaregulatory a gency ) will result fromthemanager's actions (i.e., certainty)an d howmuchofa problem thenegativeoutcome would cause for therespondent (i.e., severity). Ce rtainty (0= nochanceat all to10=100% chance)andseverity (0= noproblemat all to 10=avery big problem ) w ere m easuredon an11-point sca le.

    IV. A N A L Y S I S A N D R E S U L TS ble reports onbivariate relationships b etwee n reported offendingintentions and the key ind epend ent regu latory v ariables of interest.

    Contradic tory f indings may em erge f rom this var iable . Consis tent wi th deterrence ,therisk of offending shouldbe low if law isperceivedto bepun i t ive . However , proceduraljus t ice anddefiance theory w ould predict greater offending risk if lawisperceived to be overly strict , since thismay tap into perceptions thatlaw isunfair andillegitim ate. SeeTO M R. TYLER, W H Y PEOPLE OBEY THE LAW(1993); Lawrence W. Sherman, Defiance,Deterrence and Irrelevance: A Theory of the Criminal Sanction 30 J. RES .CRIMEDELINQ.445 (1993).

    The constructionof these scalesisdiscussedinmore deta il(as areother variables)inAppend ixIV.'^ This excludes regulatory investigation, which only has direct implicationsforthe f i rm .We calcula ted the scale in the following m anner: Individu al Formal San ction Risk =(Certainty of cr iminal '*Severity of cr iminal) + (Certainty of civil Severity of civil).Certainty= thecertaintyof outcomes [arrest (criminal); being personally sued (civil)];and

  • 8/12/2019 Corporate Environmental Crime

    25/49

    2 3] CORPOR TE CRIME CONTROL 255Table2

    Correlation Matrix

    r ^ t n S t ^ ^ o o O ' O ' ' ^ ^ c o ' ^ o o t Nv t n ^ ' ^ ^ - ' d f * ^ ' ~ t ^ ' ' O t ' '^ p p p p p p - ' p p p S ' - ^ t N f n o 00 ^ N * * f f * c^^ O ^ n v O f n O ^ O f S m ^^ , ^H ^5 (^ ^^ ^^ ^H 1 1 ^5 ^^ s ?oo o-o o o o o o o o o o o o o o

    ^ ^* n o o o i n o m o o ^ -' - i n i > - ^ ^ p ' i ' o ^ op ^ p p r n p c s f o

    O O O 00(S2S

    00 es ^ ^i n ^ ^ ' ^ i n o ^ - ^ o ^ c ^ O N C S O o o t i n^ p p p p p p p p p p p p p p p p p o o o o

    ^ p p p p p p p p p p - ^ p p p p p O O

    5 g 2 2 s n o^ (Sg st^ Ov o^ m P o o oo o o o o o o o o o o o o o

    s o t ^ i n o v - ^ o m ^ v o ^ s o i n f - p n r ^ s^ f T ) , h . - o ^ r ^ ' O (S c s u ^ f ^ s ^ - i n S q q q q S q S q S q - o q - q q q O ts 2o

    vo vfn voO .^r o v o v T r v o ^ p ^ o nO O O O O O O . 65O O O O O O O O O o o o o o o o o

    (SO5s go o o o o^ ^ ^^ r* H c^ r* * ^ ^^ ^j ^ ^ ^ ' t iin

    -2 gd d

    o 2S n r'i -^ n r^ ^ Oo r n ' - ' O ^ p p p p p p p p p - ^ p r s nd d d d d d d d d d o o o o o o o o o o on o n Cv o ^

    ^ ,-H p p ^ m * 0 ^ o t N O N O < N O t N > n ^ t N - a \ o o r - -p p p c S p p p r s p p p p p p p p m - f c s - H

    f*iooi n - ^ ^ o o m ^ - ^ ^ o o c s ^ O r o o ^ ^ ( o ( N j o O O ^ ^ O ~ O O O p P ^ ^ p * ~ ' P ' ~ p ~ |' d d d d d d d d d d d d o d d d d d (N p (N

  • 8/12/2019 Corporate Environmental Crime

    26/49

    6 S YS SIMPSON ETAL [Vol 103

    o T f ^ (N-J

    s = C3 o o

    0 00 0o

    __

    o

    o

    O

    soo

    T fMOo

    0 0go

    o

    s0 0oo

    o

    O \00p

    oo

    o

    c

    0 (N r r ^ r-- (N 1 c J

    o i/^ 00 t~- (N I p pO O O O

    ^ o o o o r ^ o o N / ^ t n i ^ r - ' ^ i n ^ o r - ' - ^ ^ r - o o ^ ^ p p p p p p p p p o o

    O N O O O f S ( N O O f n r - - 0 0 p p p p p p p p p

    II M

  • 8/12/2019 Corporate Environmental Crime

    27/49

    2013] CORPORATE CRIME CONTROL 257Several of the relationships were significant and in the predicteddirection. For instance , informal sanctions were negatively related tomanagerial intentions (offending propensity, r^i = -.386, p < .01).Managers who perceived substantial discovery costs were less apt to violatethe law. Similarly, intentions were inhibited when managers were depictedas being reprimanded within the company for engaging in similar acts ( 0 =-.137, p < .05) and when formal sanctions were perceived as likely {rpb =-.280, p < .0 0 1 ) . Conversely, reported offending w as more likely whencompanies had mandatory ethics training as part of an intemal compliancesystem (O = .134, p < .05). As responde nts' assessments of environmental

    law moved toward overly harsh or punitive, crime propensity increasedslightly{rph= -.124, p < .10), suggesting a potential defiance effect.Our bivariate correlations also revealed significant relationshipsbetween risk factors, especially those measured at the individual level, andoffending propensity. Illegal intentions increased when the act was though tto be thrilling (r^* = .2 90 , p< .0 0 1 ) or likely to bring career benefitsrpA= .352, p < .0 0 1 ). Consistent w ith our social-control argument,offending appears more likely when belief in the moral authority ofthe lawis variable (r^j = .163 , p < .01). M ost situational risk factors that capture

    firm-level processes (e.g., management level, economic constraints, andforeign competition) were not significantly correlated with offending withthe exception of our measure of authority structure. Reported offending ismore likely under the condition of a supervisor's request (i.e., asked byon e's supervisor to act illegally, = .142, p < .05).As previously noted, our dependent variable is a dichotomous outcomeso logistic regression is used to assess variable relationships. To enhancestatistical power, the two scenarios are combined and analyzed together.Therefore, the majority of people contribute two responses to each set of

    analyses and the observations are not independent of one another, which isa key assumption of multivariate regression. When this assumption isviolated, coefficient estimates will be consistent (i.e., unbiased), butstandard errors are no longer va lid . For this reason, we estimate robuststandard errors using the Huber/White/sandwich estimator to take intoaccount the lack of independence am ong ob servations.

    ^ We considered whether combining the outcomes made sense analytically as well asstatistically. Logically, it seems reasonab le to couple the illegal behav iors since both violatelegal requireme nts or standards, albeit by different degree s. W e statistically confrol foroffense type in our models.

    SOPHIA RABE-HESK ETH & ANDER S SKRONDAL, MULTILEVEL AND LONGITUDINALMODELING USING STATA 34 2005).

    ' An altema tive method for hand ling lack of indepen dence betwee n observations is toestimate a random effects model, which allows the intercept to vary across individuals.

  • 8/12/2019 Corporate Environmental Crime

    28/49

    258 SALLY S SIMPSON ET AL [Vol.103Our logistic regression analysis proceeds in stag es. In the firststage, we examine the effect of firm and individual risk factors onoffending intentions (including control variables). In the next mode l, weadd the effects of informal sanc tions. We know from the social control anddeterrence literatures that informal sanctions may operate independently toinhibit offending, but informal sanctions may also be triggered as aconsequence of intemal compliance structures or by command-and-controlinterventions. Therefore, it is important first to examine informal sanctionsseparately from other regulatory elements. Our third model includescompany self-regulation variables withou t informal sanctions. This enables

    us both to assess whether elements of this control strategy affectwillingness to violate and to examine if the effects of the risk factors aremitigated by the addition of these variables. Our fourth model com binescom pany self-regulation and informal sanctions. The fifth modelsubstitutes the command-and-control measures for the self-regulation andinformal sanctions variables. Finally, we estimate a full model includingboth firm self-regulation and command-and-control measures to assess howthese strategies operate simultaneously.' ^ For the sake of brev ity, we have

    While estimating robust standard errors treats the correlation among time-varying variablesas a nuisance, random effects models explicitiy model the lack of dependence anddecompose the total residual into between- and within-individual components. Id at 74.Given that otir research question does not require us to estimate the size of the between- andwithin-person error terms, we chose to use th simpler statistical method that requires theestimation of fewer parameters.To assess whether we could combine data from the two samples, we selectedvariables that were significant in the regression models for the merged sample andconducted se parate analyses by sample source. As shown in App endix V, there are notabledifferences betwe en samp les. For instance, offending intentions are decreased for TM onerespondents for the major pollution event (compared with defying an EPA order) and wheninformal san ction threats are high. Intentions increase when the act is perceived as thrilling.Among NU respondents, intentions are lessened when managers perceive the risk of formalsanctions to be high and consequential but increased when the depicted behavior is viewedas desirable and beneficial to the resp ond ent's career. Although different variables aresignificant for the two samples, the signs for the coefficients are similar across samples andresults are consistent with extant literature. Therefore, we have decided to merge thesamples to enhance statistical power. Future research would benefit from ex ploring how riskand protective factors as well as the success of prevention-and-control strategies may varyby compan y or industry characteristics. That particular question is beyond the scope of thisresearch.

    ' ^ We conduct a total of 145 hypothesis tests (including controls). Given this largenum ber of tests, we would expect seven significant results to occur by chance alone . Ourresults, however, reveal forty-nine significant relationships (including controls)substantially more than would be expected by chance. M oreover, these findings are

  • 8/12/2019 Corporate Environmental Crime

    29/49

    2013] CORPOR TE CRIME CONTROL 259removed the control variables from the Tables and report those results inAp pen dix VI.' ^ Table 3Regression of Behavioral Intentions on Risk Factors, Regu latoryCom ponents, and Relevant Control Variables' (N 237)

    Constant

    Model 1:Risk FactorsB (Robust SE)

    Odds ratio1.21 (.92)

    Model 2:Informal SanctionsB (Robust SE)

    Odds ratio0.23(1.00)

    Model 3:Self-RegulationB (Robust SE)

    Odds ratio2.67 (1.34)**Situational RiskAsked

    Midlevel managerEconotnicallydeterioratingForeign competition

    .51 (.33)1.66-.12 (.35)0.89-.40 (.47)0.67-.51 (.39)0.60

    .53 (.38)1.70-.05 (.37)0.95-.41 (.50)0.67-.48 (.41)0.62

    .46 (.34)1.58-.11 (.35)0.90-.37 (.47)0.69-.57 (.43)0.57Individual Risk

    Advance career'ThrillConditional compliance

    .23 (.07)***1.26.43 (.15)***1.53.05 (.06)1.05

    .18 (.08)**1.2049 (.14)***1.64.04 (.07)1.04

    .24 (.08)***1.27.46 (.14)***1.58.05 (.06)1.05Seif-regulationVoluntary reduction

    Ethics guidemanagementEthics are distinct

    -.23 (.39)0.79-.51 (.49)0.60-.13 (.47)0.88

    ' ^ Several of our co ntrol variables are significantly associated with w illingness toviolate envirotimen tal regulations. Individuals are less willing to violate regulations whenthe noncompliance is of a more significant nature (i.e., releasing toxins vs. ignoring acompliance order) and when they perceive that there is a greater likelihood that the behaviorwill endanger hum an life or wildlife. They are more w illing to violate when they view thebehav ior as mo re desirable. In addition, individuals who report more busines s experienceare marginally less willing to violate environmental law. It is important to note, howev er,that several control variables have no effect in the mo del. Scenario realism does not affectoffending intentions, nor does sample origin (NU) or personal experience/knowledge of actsdepicted in scen arios. This latter null finding, which is consistent across all of our mod els, issurprising considering the argum ent that decisions from experienc e and decisions fromdescription can lead to dram atically different choice beh avio r. Ralph Hertw ig et al..Decisions from Experience and the Effect of Rare Events in Risky Choice, 15PSYCHOL. SCI.

  • 8/12/2019 Corporate Environmental Crime

    30/49

    260 SALLYS SIMPSON ETAL [Vol. 103Random auditsSelf-reportingEthics trainingHotlineFiredReprimandInformal sanctions^

    Command and ControlFormal sanctionsAdequacy of laws

    -.96 (.25)***0.38

    -.47 (.62)0.63-.33 (.61)0.72.15 (.53)1.16-.05 (.65)0.95-.54 (.45)0.58-.70 (.51)0.50

    nter ctionsCareer* informalsanctions^Pscudo r .27 .33 .29

    Continued

    Constant

    Model 4:Sei tReeulation andInformalB (Robust SE)

    Odds ratio1.19 1.41

    Commandand ControlB (Robust SE)

    Odds ratio.37(1.10)

    Model 6:Full Model

    B (Robust SE)Odds ratio1.33 1.54

    Model 7:Interactions

    B (Robust SE)Odds ratio1.50 1.58Situational RiskAsked

    Midlevel managerEconomicallydeterioratingForeign competition

    .47 (.39)1.61-,03 (.38)0.97-.37 (.52)0.69-,48 (,45)0.62

    .67 (.35)*1.95-.06 (.38)0.94-.25 (.49)0.78-.43 (.40)0.65

    .61 (.41)1.83-.04 (.40)0.96-.22 (.52)0.81-.43 (.45)0.65

    .62 (.41)1.86-.06 (.39)0.94-.16 (.52)0.85-.36 (.45)0.70Individual RiskAdvance career^ThrillConditional compliance

    .19 (.08)**1.21.51 (.13)***1.67.04 (.06)1.04

    .24 (.07)***1.27.53 (.17)***1.70.04 (.06)1.04

    .20 (.08)**1.22.57 (.15)***1.76.03 (.07)1.03

    .26 (.08)***1.30.54 (.15)***1,72.03 (.07)1.03Seif regulationVoluntary reduction

    Ethics guide-.03 (.42)0.97-.43 (.49)

    11(.45)0.90-.39 (.51)0.01 (.45)1.01-.37 (.52)

  • 8/12/2019 Corporate Environmental Crime

    31/49

    2013] CORPORATE CRIME CONTROL 261Ethics are distinctRandom auditsSelf-reportingEthics trainingHotlineFiredReprimandInformal sanctions'

    -.22 (.49)0.80-.24 (.64)0.78-.02 (.62)0.98.21 (.56)1.23.00 (.71)1.00-.41 (.48)0.66-.60 (.52)

    0.55-.91 (.26)***0.40

    -.16 (.53)0.85-.34 (.65)0.71-.32 (.63)0.73.13 (.60)1.13-.11 (.72)0.89-.50 (.51)0.61-.66 (.54)

    0.52-.72 (.27)***0.49

    -.20 (.54)0.82-.42 (.64)0.66-.35 (.63)0.71-.35 (.63)1.09-.03 (.72)0.97-.40 (.50)0.67-.59 (.54)

    0.55-.98 (.32)***0.37 ommand and ontrol

    Formal sanctionsAdequacy of laws

    -.77 (.23)***0.46-.04 (.11)0.96

    -.56 (.27)**0.57-.07 (.11)0.94

    -.63 (.28)**0.53-.05 (.12)0.95nter ctionsCareer* informal

    sanctions'Pseudo r .3 5 .3 1 .36-.12 (.06)*

    1.13.37Note: *p < .10,**p

  • 8/12/2019 Corporate Environmental Crime

    32/49

    262 SALLYS. SIMPSON ETAL. [Vol. 103sanctions are included in the model.' '' The two variables are modestlycorrelated (r = .315); thus, it appears our respondents believe informal costsare higher for acts they perceive as more dangerous.

    Model 3 includes our variables capturing elements of company selfregulation . None of the self-regulation variables significantly affectoffending intentions, nor do they mitigate the effect of the individual- andfirm-level risk factors. Directionally, however, most of the variablesoperate in a predictable manner (except for mandatory ethics training andan anonymous hotline). Notice that the inhibitory effect of danger to life onoffending intentions becomes significant once again in this model.' ^Next, we examine the effect of including firm self-regulation variablesand perceived informal sanctions (model 4). The results are substantivelythe same as the previous two models, with the firm self-regulation variablesremaining nonsignificant while informal sanctions exert a stronglysignificant infiuence (odds of being willing to offend decrease 60% forevery unit increase in perceived informal costs). The previously significantrisk and control variables remain so in model 4 with the exception ofperceived danger, which again is rendered nonsignificant by the inclusionof informal sanctions.' ^In model 5, we examine variables capturing command-and-controlregulatory techniques, controlling for individual- and firm-level risk factors.Individuals who believe that they will face more certain and severe formalsanctions are significantly less willing to violate environmentalregulations.' '' A one-unit increase in the formal sanctions costs scaledecreases the probability that the respondent is willing to violateenvironmental regulations by about 54% . In contrast, the perceivedadequacy of the law governing the violating behavior is unrelated to

    willingness to violate the law. Several ofth individual- and firm-level riskfactors remain significant in this model. For instance, neither careeradvancement nor thrill ofth act is mitigated by formal sanction risk in thismodel. How ever, this is the first model in which being asked to offend by

    ' ^ See infraAppendix VI.' ' See infraAppendix VI. See infraAppendix V l. We also exam ined the effects of formal sanctions directed at the firm (criminal, civil,and regulatory sanction certainty and severity), but found that they were highly collinearwith the individual-level formal sanctions variable. When both scales were included in themo del, neither achieved statistical significance. When entering the variables separately,individual-level sanctions exhibited a stronger and more consistent effect on offending(refiecting the notion that managers may be more concemed with formal sanctions directedat them selves) than did firm-level formal sanc tions. Therefore, the decision was mad e to

  • 8/12/2019 Corporate Environmental Crime

    33/49

    2013] CORPORA TE CRIME CONTROL 263on e's supervisor increases the propensity likelihood. Although the odds ofoffending are 95 higher for those who have been asked to offend by asuperior versus those who have not, the effect is only marginally significant p

  • 8/12/2019 Corporate Environmental Crime

    34/49

    264 SALLYS. SIMPSON ETAL. [Vol. 103One possible explanation for the continued importance of perceived

    career benefits on offending is that the tme relationship between regulatorylevers, benefits, and noncompliance is multiplicative instead of additive.So we explored some likely interactions between variables drawing fromthe extant literature. W e expected, for instance, that people who stronglybelieved the illegal act would benefit their careers would be less likely to bedeterred by potential formal legal pro ceed ing s' ' and informal sanctionthreats. Further analysis (not reported here) failed to reveal a significantinteraction between formal sanctions and perceived benefits to one's career,but we did discover a m odest interaction (p < .07) between career benefitand informal sanctions {seeTable 3, model 7). For people who perceivegreater career advancement associated with the illegal act (compared withthose who perceive less benefit), informal sanctions matter less inpredicting offending likelihood. Using predicted probabilities todemonstrate, when the person perceives no career benefit, the probability ofoffending decreases by 0.25 (from 0.31 to 0.06) as the perceived informalsanctions increase from 1 SD below the mean to 1 SD above. When theperson ranks the likelihood that offending would advance his career as a 5(out of a possible 10 ),' the probability of offending only decreases by0.171 (from 0.47 to 0.30) when perceived informal sanctions are increasedfrom SD below the mean to SD above. This implies that when a personperceives a large career benefit, she is less likely to consider infomialsanctions before deciding to offend. The perceived benefit of illegalbehavior for this group appears to tmmp any anticipated loss of respect andfuture harm to job prospects associated with the informal discovery thatprom otes crime inhibition for others in the sample. Although this finding ismodest, ^ it points out that some regulatory elements may be less salientfor m anagers who are more instmmentally oriented.

    The empirical literature also suggests that regulatory interventionsmight operate differently for experienced responden ts. W e ran models 1 -6 on a subset of scenarios in which the respondent reported personalexperience with the environmental conditions described in any of the

    Both types of formal sanctions (threats directed at the firm and individual mana gers)were analyzed for multiplicative effects. To ease interpretation, the variables were mean-centered prior to creating theinteraction. M ean-centered value s are reported in mode l 7. Non e of the firm self-regulatoryvariables (e.g., elements and operation of an intemal compliance system) interacted withcareer benefit.' Only 13%of the sample ranked career benefits above a 5. ^ Model comparisons (with and without the interaction term) fail to show an improvedfit for the model w ith the interaction. How ever, the pseudo r^ for the mode l increases fi-om

  • 8/12/2019 Corporate Environmental Crime

    35/49

    2013] CORPORATE CRIME CONTROL 265scenarios (N = 177). W hile some coefficients dropped to nonsignificancein these models (possibly due to the reduction in sample size), the overallresults were similar. There were not enough cases to run all six models onthe subsample of scenarios in which respondents reported no personalexperience (N = 60), but a likelihood-ratio test indicated that models 1-3were not significantly different for respondents with personal experienceand those w ithout.

    V. IMPLICATIONS FOR REGU LATOR Y STRATEGYThis research was conducted to leam more about corporate

    environmental crime prevention-and-control strategies in the context of thekinds of pushes and pulls toward crime that company managers mayexperience. Specifically, we exam ined whether offending intentions wouldbe lessened in the presence of particular regulatory elements drawn fromcooperative and punitive intervention strategies. Although our sample ofrespondents is nonrandom, a nontrivial number of managers in this studyreported a chance of offending under experimentally generated conditions(N = 61 , or 'il.9 of the respondents). For this select group of managers,certain crime prevention-and-control strategies are more successfiil thanothers. For instance, the perceived certainty and severity of legal sanctionsthat target responsible managers deter environmental wrongdo ing. Thu s, asothers have discovered , ^ credible enforcement