anonymous shell companies: a global audit study and field ... · anonymous shell companies: a...

Anonymous shell companies: A global audit

study and field experiment in 176 countries

Brent B Allred1,Michael G Findley2,Daniel Nielson3 andJ C Sharman4

1Raymond A. Mason School of Business,

The College of William & Mary, Williamsburg,VA, USA; 2Department of Government, University

of Texas at Austin, Austin, TX, USA; 3Department

of Political Science, Brigham Young University,Provo, UT, USA; 4Department of Politics and

International Studies, University of Cambridge,

Cambridge, UK

Correspondence:BB Allred, Raymond A. Mason School ofBusiness, The College of William & Mary,Williamsburg, VA, USAe-mail: [email protected]

AbstractTo test whether firms behave consistently with international law prohibiting

anonymous incorporation, we conducted a global audit study and fieldexperiment, using data from 1639 incorporation firms in 176 countries. We

requested anonymous incorporation and randomly assigned references to

international law, threat of penalties, norms of appropriate behavior, or aplacebo. We find a substantial number of firms willing to flout international

standards and show that those in OECD countries proved significantly less

compliant with rules than in developing countries or tax havens. Firms in taxhavens displayed significantly greater compliance and were sensitive to

experimental interventions invoking international law.

Journal of International Business Studies (2017) 48, 596–619.doi:10.1057/s41267-016-0047-7

Keywords: shell corporations; international law; audit study; global field experiment; taxhavens

INTRODUCTIONIn the fall of 2014, Russian nuclear executive Vadim Mikerin wasdecoyed by a fake car alarm into leaving his Washington, DC-areaoffice to be whisked away by US law-enforcement agents. They wereattempting to coerce Mikerin to serve as an undercover mole in aprobe of suspected kickbacks and bribery involving the sale ofRussian uranium in the US (Schectman, 2015). Mikerin was takento a ‘‘war room’’ equipped with recording devices and radiotransmitters and festooned with diagrams of his alleged dealingswith Russian nuclear company executives through shell corpora-tions. Mikerin was suspected of participating in schemes tofacilitate and hide bribes and kickbacks paid to various officialsby sending them through shell companies and secret bankaccounts in Latvia, Switzerland, and Cyprus.

While this example may seem as if it were lifted from the plot of aspy novel, the true story underscores that the use of shellcorporations often enables illicit and nefarious activities. Althoughshell companies can serve legitimate legal purposes, they haveoften been associated with bribery, money laundering, tax evasion,and drug trafficking.1 The use of shell companies created headlinesin early 2016 when the Panama Papers scandal revealed howextensively they were being employed by heads of state and thewealthy.2 Because of these problems, the question arises as to

Received: 31 August 2015Revised: 20 July 2016Accepted: 22 September 2016Online publication date: 12 January 2017

Journal of International Business Studies (2017) 48, 596–619ª 2017 Academy of International Business All rights reserved 0047-2506/17

www.jibs.net

whether differences in legal institutions and finan-cial regulation cross-nationally might causechanges in incorporation patterns, which is thefirst node in corporate governance, and whetherthose domestic institutions might also conditionthe effects of international standards on firmbehavior.

This moves the focus of concern to globalcorporate governance, where there is a need tobetter understand the ‘‘conflict, convergence, coop-eration and competition among legal systems,institutions, actors and rules’’ (Wood, Abbott,Black, Eberlein, & Meidinger, 2015: 334). Globalcorporate governance adds extra layers of complex-ity to single-country contexts and has thus requiredthe creation, proliferation, and stabilization of newgovernance systems transnationally (Djelic &Quack, 2003). Sugathan & George (2015) foundthat the corporate governance effectiveness and thequality of country-level governance infrastructureinfluence the profit shifting of foreign firms out ofhost countries. Globerman & Shapiro (2003)reported that countries that achieve effective gov-ernance are more likely to receive foreign directinvestment from the United States. Hiller et al.(2011) related evidence that various country-levelgovernance mechanisms significantly affect R&Dinvestment including strong law enforcement,effective board control, and a bank-based financialsystem.

Many years ago, Hoskisson, Eden, Lau, & Wright(2000) identified domestic legal, financial, andpolitical institutions as critical in influencingstrategies that transnational businesses adopt. Sincethen, a consensus has developed among interna-tional management scholars that institutions mustbe considered alongside transaction costs,resources, and corporate governance/agency theory(Kostova, Roth, & Dacin, 2008; Meyer, Estrin,Bhaumik, & Peng, 2009; Peng, Wang, & Jiang,2008; Wright, Filatotchev, Hoskisson, & Peng,2005). Of course, the assertion that institutionsmatter does not provoke controversy; the keyquestion is how and why they matter, especiallyacross varying national contexts (see Peng et al.,2008).

Individuals and companies must navigate bothformal and informal institutions as they do busi-ness abroad. Formal institutions include the legal,taxation, and developmental policies of a country(Redding, 2005). Informal institutions are networksand nested organizations influenced by local cul-ture and accepted norms (Leung, Bhagat, Buchan,

Erez, & Gibson, 2005). Institutions therefore pro-vide the context, or the formal rules and informalnorms, that organize and govern individual, non-governmental organizational, and business interac-tions (Teegen, Doh, & Vachani, 2004). Domesticinstitutions in some countries may encourage orfacilitate practices (e.g., anonymous incorporation)that the domestic institutions of other countriesmay deter or prohibit altogether. Scott (1995: 33)thus describes institutions as the ‘‘regulative, nor-mative, and cognitive structures and activities’’ thatcreate meaning and stability for actions and behav-iors. Therefore institutions broadly define societyby forming, in the words of Douglas North, ‘‘therules of the game’’ (Meyer & Peng, 2005; North,1990). It is within institutional constraints – bothformal and informal – that firms pursue theirinterests (Ingram & Silverman, 2002).However, there is a general consensus in inter-

national business studies that too little is knownabout how institutional contexts differ cross-na-tionally and how variance in context channelsconstrains business activities from country tocountry (Cantwell, Dunning, & Lundan, 2010;Peng et al., 2008; Redding, 2005). In an attemptto contribute to a better understanding of thevariation in institutional contexts globally, in thisarticle we categorize countries into multiple broadtypes whose formal and informal institutions dif-ferently affect how new businesses can be legallyincorporated – in ways that carry large implicationsfor firms’ compliance with international rules. Wedo this through an audit study with accompanyingfield experiment that offers strong causal inferencecoupled with high naturalism or ecological validity(see Chatterji, Findley, Jensen, Meier, & Nielson,2016).The focus of the study is the creation of shell

corporations internationally. Shell companies arelegal entities that typically do not produce sub-stantive products or services.3 They may be createdby private individuals for legitimate purposes, oftenas holding companies. However, shell corporationsmay also be used to obscure the real owner of assetsor the controlling party behind financial transac-tions. This kind of company is thus very differentfrom the typical multinational corporation of pop-ular imagination and most international businessresearch. Shell companies are almost always pri-vately held, not publicly listed.4 As largely passiveholding entities, they are indifferent to profit andloss. Because they produce no substantive good orservice, their directors and other officeholders are

A global audit study and field experiment in 176 countries Brent B Allred et al

597

Journal of International Business Studies

formally exercising legal duties on behalf of theindividual owner, not performing any practicalmanagement function. They are the first node incorporate governance as limited liability entities, incontrast to the unlimited liability of partnerships orsole proprietorships. As such, incorporation shouldbe a central issue in management research.

The challenge arises when individuals seek un-traceable incorporation, particularly internation-ally, in which the beneficial (real) owner of theshell company cannot be identified. Shell compa-nies whose real owners cannot be discovered mayenable a slew of financial crimes. In response to thisissue, the Financial Action Task Force (FATF, 2016),an international organization created by theworld’s leading economic powers, was created topromote and enforce corporate transparency stan-dards worldwide. These standards, now accepted by180 countries and instantiated in a variety ofinternational conventions and laws, prohibituntraceable shell companies by requiring thatgovernments must be able to ‘‘look through’’ suchlegal persons to find the real individual or individ-uals in control, in other words, the beneficialowner.

However, there is a demand for untraceable shellcorporations that is in direct tension with the FATFrules. Moreover, the FATF does not audit theintermediary firms – corporate service providers(CSPs) – that create shell companies in membercountries to learn if corporate transparency stan-dards are actually being followed in practice.Rather, they assess domestic laws to verify thatthey harmonize with FATF rules. This gives rise to apotential disconnect between international rules,domestic institutions through informal norms andformal laws, and firm behavior, which is at theheart of this study. Hence the crux of this article is aprecise set of tests concerning the effects of infor-mation regarding international institutions onincorporation services’ conduct, and subgroup testsof how domestic institutions condition or moder-ate firms’ actions in response to information aboutthe global rules. The subgroup tests by country typeprovide precise measurement of the effects ofdifferences in institutional context on compliancewith global rules. The subgroup tests can thusreveal variation in treatment effects induced by theexperimental interventions across different cate-gories of country institutional environments.

In this manuscript, we first consider how wellinternational FATF standards have been imple-mented at the national level with a comparison to

incorporation practices as measured in our study.In this, we demonstrate that the grades issued tocountries by the FATF in their periodic reviews donot correlate particularly well with the findings oncross-national compliance from our audit study,which is the first global assessment conducted atthe firm (corporate service provider) rather thancountry level. With this discrepancy as back-ground, we then outline theoretical ideas thatmay provide some explanation for why firms mightcomply or fail to comply with international stan-dards. These ideas include simply making firmsaware of regulations, as expected in a managerialinternational law approach, as well as highlightingthe consequences or appropriateness of certainbehaviors.For the audit study, researchers adopted personas

as putative consultants seeking ‘‘confidential’’incorporation, and 2161 requests were sent to1639 corporate service providers in 176 countries.A central goal was to uncover global levels ofcompliance with international standards. In theaudit study, we show that levels of compliance arenot high, and sometimes troublingly low. Observa-tionally, we find substantial firm-level non-compli-ance globally, and in investigating the effectswithin different country groups, we find that firmsin tax haven countries were significantly morecompliant with transparency standards comparedto CSPs in OECD countries and developing nations.This is the study’s first evidence that different typesof countries evince significantly distinct institu-tional contexts for incorporation.As part of the audit study we embedded a field

experiment by randomly assigning informationregarding FATF standards, including a promptnoting the rules on identification documents, awarning that there might be legal penalties, and aninvocation of norms of appropriate behavior. Theexperimental results indicated few treatment effectsfor the different international law informationconditions across the global sample. However,again providing strong evidence for the importanceof institutional context, while CSPs in developedand developing countries proved largely indifferentto the information prompts about standards com-pared to a neutral placebo, firms in tax havensdisplayed significantly higher sensitivity to thetreatment conditions. This suggests pronouncedcross-national heterogeneity in domestic institu-tions’ conditioning effects on firm behavior whenpotential violations of international standards areencountered. Firms in tax havens responded


598


markedly to prompts invoking international law,indicating that the institutional context for incor-poration is very different for this type of countrycompared to wealthy developed and poorer devel-oping nations.

The research design and analysis plan were pre-registered with an interdisciplinary research con-sortium (www.egap.org) prior to the study’s exe-cution, and the plan described the inclusion ofcountry type as a blocking criterion. Therefore thissubgroup analysis is prescribed as best practice inexperimental methods and should alleviate con-cerns about post hoc ‘‘fishing’’ for significantresults. The results offer suggestive evidence thatvariation across institutional forms captured bycountry type likely affects firm behavior and mod-erates the effects of varying information aboutstandards – all dynamics in need of follow-upresearch. In what follows, we articulate the contextfor the study, detail the research design, and pre-sent our results.

FORMAL LAW VERSUS ACTUAL PRACTICE

In 1989, the G7 countries established the FinancialAction Task Force (hereafter FATF) to combatmoney laundering by promoting financial trans-parency, including corporate transparency. Thisorganization developed key recommendations rec-ognized as the standard for battling money laun-dering and terrorist financing. Aside from the G7,the FATF now includes all members of the G20 anda large majority of the OECD countries. Nearlyevery other country is a member of regional FATFassociate bodies committed to upholding FATFstandards. Although the FATF itself only producessoft law, codified in 40 Recommendations, its ruleshave been explicitly endorsed by the UnitedNations Security Council, incorporated withinWorld Bank and International Monetary Fundstandards, and written into a range of internationalhard law conventions. Three of these Recommen-dations have been the most important for improv-ing corporate transparency through identificationof the beneficial owner.

The first (#5), deals with Customer Due Diligencein mandating that banks establish the identity ofall account holders, including the identity of theowners of companies that hold corporate accounts.The second (#12) imposes the same Customer DueDiligence requirements on entities such as casinos,real estate agents, lawyers, and, crucially for ourpurposes, the corporate service providers that set up

and sell shell companies. The third relevant rec-ommendation (#33), mandates that governmentsbe able to find the real (beneficial) owner of allcompanies in stating that ‘‘countries should ensurethat there is adequate, accurate and timely infor-mation on the beneficial ownership and control oflegal persons that can be obtained or accessed in atimely fashion by competent authorities’’ (FATF,2013: 22).5

While the problem of untraceable shell compa-nies has attracted renewed attention from the G7,G20, and other bodies in the wake of recentfinancial crises, efforts have been devoted to betterenforcement of the FATF’s basic rule on the trans-parency of companies first laid down in 2003,rather than instituting new standards. As a staterather than federal responsibility, incorporationstandards have been largely constant in the US.The EU and Britain introduced important newregulations on corporate transparency in 2015,but the basic goal remains meeting the FATF rulequoted above on identifying beneficial owners.While it is important for a country to have

official laws regarding specific requirements forincorporation, such laws are generally not veryuseful if they are not followed in practice (see, forexample, Allred & Park, 2007 in regards to formallaw versus practice for patent rights). Having lawson the books may help establish legitimacy withthe global community, but the lack of enforcementundermines this legitimacy and provides openingsto the type of illicit activities that such lawsattempt to discourage.To learn how well domestic laws are enforced in

practice, we performed a test of the three key FATFrecommendations comparing national-level com-pliance as assessed by the FATF to the actual firm-level practice based on our audit study. The FATFconducted Mutual Evaluation Reports (MERs) todetermine whether a country is Compliant (C),Largely Compliant (LC), Partially Compliant (PC),or Non-Compliant (NC) with each of the 40Recommendations. The MERs consisted of a com-bination of questionnaires, onsite inspections, andplenary discussions of countries’ performance.Appendix B provides a summary of the FATF MER

assessments, focusing on the three key FATF rec-ommendations discussed above. FATF MERs werecompleted for 156 countries for Recommendation5, with 72 (46.2%) of the countries being Non-Compliant, 71 (45.5%) being Partially Compliant,13 (8.3%) being Largely Compliant, and no coun-tries being fully Compliant. For Recommendation


599


http://www.egap.org

12, 162 countries have MERs data, with 114 (70.4%)of the countries being Non-Compliant, 44 (27.2%)being Partially Compliant, 4 (2.5%) being LargelyCompliant, and no countries being fully Compli-ant. Recommendation 33 had MER data for 160countries, with 49 (30.6%) of the countries beingNon-Compliant, 85 (53.1%) being Partially Com-pliant, 20 (12.5%) being Largely Compliant, and 6(3.8%) being fully Compliant.

While it is helpful to determine the level ofcompliance at the national level as measured by theFATF’s assessment of local laws on the books, thequestion arises as to whether this translates toactual compliance at the firm level among corpo-rate service providers. Until this study, however, noaudit of individual firm behavior existed to com-pare to the results of national-level assessments. Ifthe two correlate highly, the current FATF protocolmay be sufficient. Yet there is good reason tosuspect that the alignment of domestic statutorycompliance with international standards does nottell the whole story. A central goal of this study,therefore, is to conduct a firm-level audit study ofcorporate service providers to understand levels ofcompliance at the crucial locus of behavior, whichwe can then compare to the MERs report scores. Wedetail the full design of our audit study below, butfor now we point out that national compliancefrom MER reports does not significantly predictfirm-level compliance for FATF Recommendations5 and 12 (see Appendix C). In fact, it predictsvirtually none of the variance. For Recommenda-tion 33 on beneficial ownership, national compli-ance is a significant predictor (at the p\0.05 level)of CSP compliance, but less than 4% of the varianceis explained, offering weak substantive support forthe proposition that statutory law drives firm-levelbehavior across the full sample. The disconnectunderscores the importance of the firm-level auditstudy, and it also raises questions about whyvariation in firm-level compliance may occur. Wethus turn to a discussion of some possibleexplanations.

INSTITUTIONAL CONTEXT: CONSEQUENCESVERSUS APPROPRIATENESS

The lack of strong support above raises questionsregarding why some nations may commit to inter-national standards but then fail to enforce them,while others fulfill their commitments at signifi-cant cost. Key parts of the answers lead to a re-engagement with the discussion of institutional

context previewed above. While behavior in inter-national business is likely driven in large part byfirm capabilities (Barney, 1991; Wernerfelt, 1984)and industry conditions (Porter, 1980), it is thevarious countries’ institutional contexts that deli-mit both the informal and formal constraints onmanagerial actions (Peng et al., 2008) and definelegitimate business behavior (North, 1990). Cant-well et al. (2010) posit that economic growth isdependent on the establishment of more advancedinstitutions. This requires more effective policiesand regulations. For example, Demirguc-Kunt et al.(2006) found that incorporated businesses reportedfewer obstacles to growth and operations in coun-tries with developed institutions. Although govern-mental efficiency focuses on the effectiveimplementation of public policy (Galang, 2012),different institutional factors may influencewhether and how domestic firms adhere to suchpolicy.The international business literature has often

focused on two broad types of cross-national insti-tutional contexts: developed and developing/emerging (Cantwell et al., 2010; Meyer & Sinani,2009; Wang, Hong, Kafouros, & Wright, 2012).Developed contexts demonstrate reliable rule oflaw, efficient regulatory frameworks, and hightransparency/low corruption; developing/emerginginstitutional contexts evince the opposite condi-tions (Meyer & Sinani, 2009). This is a usefultypology but, as our research details below, it omitsan important category of jurisdictions that matter agreat deal for international business and financebut cannot be captured by the dichotomous cate-gorization of developed versus developing. Themissing type is tax havens, countries that span arange of socio-economic development levels frommiddle- to upper-income but that also display adistinctive set of legal institutions and normsrelevant to incorporation and international busi-ness more generally.Both formal and informal institutions character-

ize all three types of country contexts, but precisemeasurement of the effects of formal rules versusinformal norms is difficult given traditional meth-ods relying on observational data and is thereforesubject to selection bias and collinearity, amongother problems. We thus consider different exper-imental probes drawn from arguments relating tothe two broad types of institutions.Formal institutions generally connect to March

and Olsen’s (1998) logic of consequences and infor-mal institutions relate to their logic of


600


appropriateness. The logic of consequences (hereafterconsequences) posits that adherence to laws andstandards occur mainly to avoid penalties or toachieve some benefit (March & Olsen, 1998);actions are therefore based on the anticipatedoutcomes among alternatives. Consequences appliedinternationally holds that actors only obey inter-national laws and standards to avoid sanctions orcapture gains as dictated by formal rules (Abbott,Keohane, Moravcsik, Slaughter, & Snidal, 2000;Goldstein, Kahler, Keohane, & Slaughter, 2000).These formal global rules may be transposed intodomestic law and implemented via local govern-ment agencies but only have ‘‘teeth’’ when activelyenforced. As such, national laws and regulationdrive transaction costs (Buckley & Casson, 1976), sofirms should necessarily consider the penaltiesattached to violations of the formal rules, whilealso evaluating the risks associated with interna-tional political and policy uncertainty (Miller,1992).

The decision as to whether or not to comply thenbecomes a classic cost–benefit calculation, encom-passing the real and reputational costs and benefitsthat are expected to come about from specificactions (Checkel, 2001; Drezner, 2007). If thelikelihood of being caught or punished is low, orthe ultimate cost from subsequent punishments issmall, there is limited incentive for a country orfirm to follow international law. When penaltiesare minor or extensive delays in due process occuror can be created, nations may take a ‘shoot firstand ask for forgiveness later’ approach. This tacticimplies that when there are limited costs to doingso, countries may publicly promise to abide byinternational rules but then ignore these commit-ments to pursue critical economic or securityconcerns (Goldsmith & Posner, 2005).

In contrast to consequences, informal practicescharacterize the logic of appropriateness (hereafterappropriateness), which proposes that nationaladoption and adherence to international law is afunction of and response to accepted norms (March& Olsen, 1998). The structure of the internationalsystem is a social structure (Arend, 1997) and isfundamentally comprised of and based on interna-tional law.

This system is premised on binding, meaningfulagreements that are themselves critically groundedon commonly accepted norms and beliefs. It is thesocial and moral commitment to the ideals theserules and laws are based on, not penalties or

benefits, which leads countries to adopt and adhereto common standards.Rather than being the product of cost–benefit

calculations, compliance reflects shared norms ofproper conduct, internalized through socializationwithin a community and bolstered through socialapprobation or disapproval. Trust and reciprocityreduce risk and uncertainty because codes of con-duct are institutionalized (Singh, Lentz, & Nijssen,2011; Stephan & Uhlaner, 2010). The emphasis onsocial mechanisms to ensure compliance (Gra-novetter, 1985) means that when commitmentsare not met, punishments take on a more socialnature, such as exclusion from marks of prestige,ostracism, or expulsion from a group (Thompson,1996). Checkel (2001) posits that it is through sociallearning and persuasion that norms are establishedand understood and compliance is achieved. Socialconditioning helps create structures that stabilizethe control and coordination between entitiesinvolved in economic exchange (Redding, 2005).Interestingly, Coffee (2001) suggests that normsmay matter the most when laws are the weakest.In the absence of highlighting either the conse-

quences or appropriateness of legal behavior, it isexpected that a basic knowledge of the law shouldprompt compliant behavior. The managerial schoolof international law (Chayes & Chayes, 1996), here-after managerial law, maintains that most countriesfollow most international rules most of the time.This is because governments conclude that suchrules advance their interests for bureaucraticdomestic reasons as a compliance machinery isbuilt up, and because of a general sense of obliga-tion to fulfill commitments in the internationalcommunity. Conversely, non-compliance isascribed to actors being ignorant of the applicablerules, often because the rules themselves are vagueor under-specified.By shifting the attention to firms (CSPs), rather

than countries in aggregate, we have the opportu-nity to better understand what influences actualcompliance with international law. In addition tomanagerial law, the two opposing theoretical per-spectives discussed above – consequences andappropriateness – offer contrasting hypotheses forfirm behavior. According to consequences, sanctionsand punishment should have a greater influence onfirm compliance with standards. This is contrastedwith appropriateness, which proposes that adher-ence to national, industry, or cultural normsshould better explain firm willingness to comply.


601


Since our earlier test shows that the relationshipbetween national law and compliance at the firmlevel is marginal at best, shifting attention to firmsallows us to better determine whether these theo-ries, or some other factor, can explain whichcorporate service providers on average may ormay not behave consistently with national andinternational law.

AUDIT STUDY AND FIELD EXPERIMENTON COMPLIANCE

To explore firms’ compliance with laws pertainingto incorporation and transparency, we employ afirst-of-its-kind audit study with an embedded fieldexperiment, which we conducted between March2011 and July 2012. One key challenge for studyingwhether firms comply with laws comes from thesensitive nature of this topic. A naturalistic auditstudy allows us to overcome this challenge byimproving the overall ecological validity. Moreover,by including a randomized experimental compo-nent, we can consider possible explanations forcompliance and provide stronger causal inferencesthan would be otherwise possible, especially inprobing the effects of informal norms, formal rules,and information about international law. This isparticularly relevant in exploring how firms incountries evincing different institutional contextsrespond to the same experimental interventions.Using a large sample of 1639 firms drawn from 176countries further enhances the external validity.

In contrast, a survey approach, in which firms aredirectly asked whether they comply with the law,has obvious limitations with validity and, likely,reliability. With surveys, the expectation is thatfirms, or the individuals within them, when specif-ically asked in such a fashion, would indicate andacknowledge that they adhere to all aspects of thelaw, even though they may not do so in practice. Alab setting would also not be appropriate, since itwould be nearly impossible to empanel a properpool of respondents or create a setting that wouldaccurately resemble the principle players and cir-cumstances for incorporation of firms across theworld. Thus a more natural, field-based audit studyis most appropriate for the topic at hand.

In our study, we are able to create a situation thatdoes not simply mimic the behavior under consid-eration, but actually tests such behavior with therelevant real entities, specifically corporate serviceproviders across the world. This permits us todetermine whether these providers comply,

particularly when an explicit request is made toviolate the rules, and when participants are primedwith the knowledge of these standards.We further include a field experimental compo-

nent. Field experiments have seen limited use ininternational business research and related fieldsand are just beginning to attract heightened interestas a methodology for exploring complex and diffi-cult topics in areas such as development economics(Banerjee & Duflo, 2009; Duflo, Glennerster, &Kremer, 2008), organizational economics (Bandiera,Barankay, & Rasul, 2011), behavioral economics(Harrison & List, 2004; Levitt & List, 2009), strategicmanagement (Chatterji et al., 2016), and psychol-ogy (Shadish & Cook, 2009). For this field experi-ment, we consider the theoretical perspectivesdiscussed above, managerial law, consequences, andappropriateness, in designing interventions to learnwhether information, penalties, or accepted normsare key motivators for compliance behavior acrossdifferent institutional contexts.This design overcomes many of the concerns for

survey and lab experiments because the subjects arenot self-selected into the sample and are unawarethey are being investigated. This issue is particu-larly important for research in international law,since most of the studies in this field that haveemployed survey or lab experiments (Chilton &Tingley, 2013; Hyde, 2007) have been subject to theabove criticisms. What is being tested here is theactual behavior of those involved with the incor-poration of firms. The scope of this study also givesit high external validity due to an extensive datasetthat draws on firms from 176 countries from alleconomic regions throughout the world. Suchcoverage overcomes concerns regarding global gen-eralizability from a small sample, or one that islimited in coverage of countries or geographicregions. Such broad country representation alsoallows the sample to be grouped into categories ofdeveloping countries, tax havens, or OECD coun-tries in a way that enables the exploration of effectsin different institutional contexts. Using well-established classifications means that the resultscan be more confidently generalized within andacross these categories and do not need to beextrapolated from unmatched samples.

METHODOLOGYAs noted, we employed a global audit study with anembedded field experiment. Birkinshaw, Bresman,& Nobel (2010) noted the opportunity, if not need,


602


to employ experimental approaches to interna-tional business research. They indicate that as of2010, no field experiment had been reported inJIBS, and there have been very few, if any, since. Toconduct the audit-based experiment, our research-ers employed aliases and posed as internationalconsultants desiring to create a confidential shellcorporation in the target country hosting theincorporation service.

The study thus employed deception, an ethicallyvexed and contentious approach. Under the Bel-mont Report (United States, 1978) standards, decep-tion can only be justified if the costs are low, thebenefits high, subjects experience no harm or pain,and there is no other way to conduct the study. Allconditions hold for the present research. On aver-age, subjects spent 5–10 min responding to inqui-ries, often with canned language used incommunication with many other customers, socosts were low. No firm-level data existed before thestudy and therefore the scientific and policy ben-efits of obtaining firm-level information arguablyjustified the small amount of time invested by eachof the providers. Because the profit-driven crimesenabled by untraceable shell companies compriseserious policy problems that cause a great deal ofharm and human suffering, the social benefits oflearning from the results are high. The data werefully anonymized so that no adverse consequenceswould result to the subjects, and all subjectsanswered inquiries in line with their normal day-to-day routines. Conversely, surveys or other meth-ods using fully informed consent would likely haveproduced biased results, suggesting that theresearch could not have been reliably performedusing alternative methods (see Findley, Nielson, &Sharman, 2015). The study aligns with prominentarticles in social science employing similar meth-ods (Bertrand & Mullainathan, 2004; Butler &Broockman, 2011).

We collected information on a large pool ofincorporation services for our sample, and thenconducted an audit study with random assignmentof placebo or treatment conditions to determinecompliance or lack thereof. For the placebo andtreatment conditions (discussed below), varyinginformation concerning international law regard-ing incorporation and its motivation was randomlyassigned across the pool. The primary requestacross all conditions from our fictitious individualclients, including the placebo, was for the CSP toprovide its services to create a new corporate entitythat would limit tax and legal liability and

maintain confidentiality. A request for limitedliability, tax savings, and confidentiality does notviolate international standards, only an explicitrequest to avoid supplying ID documents, or theincorporation service provider’s failure to requiresuch documentation, violates these standards.Compliance was determined at two levels. First,did the incorporation firm respond to our email?Those that did not were adjudged as exhibiting softcompliance, though as discussed further below,conceivably non-response could have been theproduct of inattention or a commercial logic.Second, did the firms that responded require thenecessary certified identifying photo documenta-tion consistent with international standards (hardcompliance for those that did respond)? Because allof our inquiries, including the placebo, asked whatidentity documents, if any, were required, thedetermination of whether providers adhered toidentification requirements in line with this querycomprised the audit portion of the study.The experimental component involved ran-

domly assigning three treatments (FATF, conse-quences, and appropriateness), which were testedagainst a control (the placebo). A reference tointernational standards alone appears in the firsttreatment (FATF), but then to test the influence offormal institutions/consequences and informalnorms/appropriateness, we added (not replaced)language to the international standards referencethat raised the possibility of consequences or theappropriateness of various norms. These threeexperimental conditions thus map to the threetheories discussed above – managerial law, conse-quences, and appropriateness – and provided infor-mation about international standards, as well asincentives or norms to follow those standards. It ispossible that some firms already understood thestandards, which would make the treatment some-thing more of a prime rather than a pure informa-tion treatment. Following the audit study, weconducted a survey and learned that roughly 70%of firms were not aware of the global standards, butthat more firms in tax havens were aware of thesestandards than those in non-tax havens (Findley,Nielson, & Sharman, 2014).Corporate service providers were categorized as

residing in OECD, Tax Haven, or Developingcountries according to World Bank and OECDclassifications. This design allowed for the testingof whether any of the treatments resulted insignificantly different compliance (soft or hard)compared to the placebo, and whether the


603


institutional context of the country in which theincorporation firm was based influenced the rate ofcompliance.

To better create a neutral basis for the nationalsource of the inquiry, all email requests came fromwealthy, low-corruption OECD country aliases. Forthis condition, the consultant claimed to be fromone of a set of wealthy countries (Norway, Finland,Sweden, Denmark, the Netherlands, Austria, NewZealand, or Australia), which we dubbed the‘‘Norstralia’’ countries for ease of reference. TheNorstralia countries are among the least corruptcountries ranked on the Transparency Interna-tional Corruption Perceptions Index (CPI). An aliasfor each country was created based on the mostcommon male names in that country.

PlaceboThe placebo condition was a basic request forconfidential incorporation, without any referenceto international laws pertaining to incorporationand transparency. In this treatment, the consultantspecifically inquired only about what documenta-tion would be required to form a corporation (seeAppendix D for examples of the placebo andtreatment emails and Figure 1 for a graphicalrepresentation of the Treatment Hypotheses).

FATF (Information on Law)To test the regulatory influence of the relevantinternational standards promulgated by the FATF,the first treatment identified the FATF and perti-nent international standards. This treatment spec-ifying the requirement for identity documents wasintended to evaluate whether the basic knowledgeof the law’s existence and requirements influenceswhether incorporation firms were more likely tocomply. This intervention thus tests the effects

on compliance of information about the formalrules. The consultant indicated that he still desiredto limit disclosure and maintain confidentialityand, as with the placebo condition, specificallyinquired about the documentation required forincorporation.As noted, the managerial school of international

law (Chayes & Chayes, 1996) posits that mostcountries follow international rules through a senseof responsibility to satisfy commitments to thebroader international community, andbecause theirinterests are best served by so doing. The failure tocomply is attributed to lack of knowledge or under-standing of relevant rules. This theory leads to thehypothesis that when given information aboutinternational standards, incorporation services aremore likely to comply with them than when theyhave not been primed with such information.

Hypothesis 1: Incorporation firms are morelikely to comply with international standardswhen informed of such standards relative towhen they are not informed.

Consequences (Formal Penalties)The second treatment mentioned international lawand raised issues relevant to consequences (Goldsteinet al., 2000; March & Olsen, 1998). As with theInternational Standards treatment, this treatmentprovides identical information about FATF and itsrecommendations for documentation when form-ing a corporation. To prime the consequences mech-anism, this treatment also added that the violationof these standards may result in legal penalties. Theconsultant then subtly indicated a desire to violateinternational law by saying that he would prefer toavoid the disclosure of personal information,implying confidentiality and non-compliance withthe law. The purpose of this treatment is todetermine how incorporation firms react whenprimed with consequences of violating the formalinternational standards and law, particularlywhether they are more or less likely to complywith the law and risk sanctions. Based uponconsequences, the expectation is that the threatand awareness of punishment would increase therate of compliance with international law, relativeto the placebo condition.

Hypothesis 2: Incorporation firms are morelikely to comply with international standardswhen informed of sanctions (consequences) forviolating such standards relative to when they arenot informed about such sanctions.

Appropriateness (Norms)

Consequences (Penalties)

FATF (Information

on Law)

International Standards

Compliance

H1 (+)

H2 (+)

H3 (+)

Figure 1 Treatment model.


604


Appropriateness (Informal Norms)The third treatment draws upon appropriateness bysuggesting that compliance with international lawbe based on norms of appropriateness and generalconformity. This treatment expressly primes thefirm by indicating that most countries are signato-ries with the FATF standards, which require specificdisclosure for incorporation. In addition, the treat-ment emphasizes that ‘‘as reputable businessmen’’both the applying firm and incorporation firmwant ‘‘to do the right thing’’ according to interna-tional standards. Even so, the consultant still asksfor the ability to incorporate by not disclosingpersonal information, effectively creating ananonymous corporation and violating the law.The appropriateness approach proposes that in orderto conform to generally accepted and sharedexpectations for suitable behavior, firms will actethically and appropriately. Thus by providingstatements about norms and acceptable practices,higher compliance is expected over both theplacebo and international law conditions.

Hypothesis 3: Incorporation firms are morelikely to comply with international standardswhen exposed to norms of appropriateness foradhering to such standards relative to when theyare not exposed to these norms.

Host Country Institutional ContextThe scope and breadth of this study’s sample allowsfor further analysis and insight in addition to thethree treatments. On top of the treatment effectsfrom the variation in the email message, we are alsoable to take into consideration host-country charac-teristics or institutional context, particularly thelevel of development and nature of the country inwhich the incorporation service is located andincorporation is being requested. This distinction isparticularly valuable since it is not expected thatfirms across the world behave consistently, as notedabove in the discussion of institutional context, butinstead that national circumstances may influencethe propensity towards compliance to both nationallaws and international standards. This allows formore nuanced tests of institutional theories ofinternational business. We divided jurisdictions intothree categories: OECD, Developing, and Tax Havencountries based on World Bank and OECD classifi-cations. Such categorization reflected the blockingcriterion of country type used to sort firms prior torandom assignment. We pre-registered this intentbefore execution of the experiment, and thus the

subgroup analysis is justified under norms of bestmethodological practice.Prior work had indicated that, contrary to popular

perception, firms in tax havens would prove signif-icantly more compliant with international law thanfirms in developing countries and, surprisingly, alsomore compliant than firms in developed countries(Sharman, 2010, 2011). Tax havens were originallyidentified by the OECD as those that host companiesare not engaged in substantive business activities,that apply low or zero tax rates, do not exchange taxinformation with other governments, and have littletransparency (OECD, 2009). Tax havens’ relativelylow power in the global system, combined with thehigh importance of the incorporation business totheir economies, made them especially vulnerableshould they appear on the FATF black list. Thesefacts led us to expect quite different behavior fromCSPs in tax havens than in developed or developingcountries. This prior scholarship suggests threeadditional hypotheses pertaining to differencesacross country groups in terms of institutions’conditioning overall compliance and sensitivity toprimes regarding international standards (see Fig-ure 2 for a graphical representation of the Hostcountry conditions Hypotheses):

Hypothesis 4: We expect CSPs in tax havens tobe more compliant with international standardsthan firms in OECD (and developing) countries(4A). Tax-haven firms should also be more sensi-tive when informed of international standards(4B), informed of the consequences for violatingthese international standards (4C), and when thenorms of appropriate behavior to follow inter-national standards are invoked (4D).

Hypothesis 5: We expect CSPs in OECD coun-tries to be less compliant than firms in tax havens(5A). OECD firms should also be less sensitivewhen informed of international standards (5B),informed of the consequences for violating theseinternational standards (5C), and when thenorms of appropriate behavior to follow inter-national standards are invoked (5D).

Hypothesis 6: We expect that CSPs in develop-ing countries will be less compliant than firms intax havens (6A). Developing country firms shouldbe less sensitive when informed of internationalstandards (6B), informed of the consequences forviolating these international standards (6C), andwhen the norms of appropriate behavior to fol-low international standards are invoked (6D).


605


SampleTo conduct this study, we needed to create a largeglobal pool of firms that incorporate companies fora fee. Information on such firms, particularly at theinternational level, has not been compiled anddoes not exist in any central database. As a result,we were required to employ a non-random sam-pling approach that drew from Internet sources andgovernmental data. Some CSPs are well-establishedcompanies or law firms that specialize in incorpo-ration services, while others provide this functionas part of their general offerings. Although we wereable to procure data from various commerciallistings and governmental sources, our approachdrew heavily on Internet searches.

Given the possibility of the existence of reclusiveincorporation firms and due to the size of thispotential market, it was impossible to obtain infor-mation on all CSPs. That said, we are confident wehave captured a meaningful percentage of theaccessible incorporation firms. Our sample selectionprocess generated firms that have a greater publicpresence than those that went undetected, whetherintentionally or not. Authorities are therefore morelikely to scrutinize the firms in our sample, which inturn are expected to have greater compliance thanlower-profile incorporation firms. Although thismay have resulted in a selection bias towards firmsthat favor compliance, such a bias would also resultin a more conservative test (and under-report non-compliance with international law).

Blocking and Random AssignmentFor the experiment, prior to the random assign-ment of firms into control and treatment

categories, we employed a blocking technique toimprove the sample design. Blocking is a proce-dure that places the sampled firms into naturallyor logically similar groupings based on covariatevalues. As a result, the experimental design allowsfor results that compare firms with similar charac-teristics and therefore helps neutralize potentialconfounds. Block randomization was thus per-formed within country categories: OECD, TaxHaven, or Developing. We also blocked based onwhether the firm was a stand-alone incorporationservice or a law firm. Blocking ensures covariatebalance across experimental conditions andremoves the possibility of high collinearitybetween covariates (country or firm type) andassignment to the various treatments or placebo(Gerber & Green, 2012).We then randomly assigned a Norstralia alias (as

mentioned, a putative consultant from Norway,Finland, Sweden, Denmark, the Netherlands, Aus-tria, New Zealand, or Australia) to each email, withan appropriately localized individual name tied tothe home country. The text and the subject linewere also randomized. The randomization of thesubject line was to help mitigate the risk ofdetection, should two associated firms receivemultiple requests. The text of the email representedthe placebo or one of the treatment conditions.Unique numbers were assigned to each treatmentcondition, alias, and email text. We then usedthe numbers to randomly assign the conditionswithin each block. Individual treatment texts orsubjects lines were not significantly associated withoutcomes, and difference-in-means tests indicatedthat there was good blocking covariate balance

Developing Countries

TaxHavens

OECD Countries

International Standards

Compliance

H4A (+)

H5A (-)

H4B (+)

H4C (+)

H5B (-) H5C (-)

H6A (-)H6B (-)

H6C (-)

Tax Havens are more compliant than OECD and Developing Countries (H4A/H5A/H6A)Treatment Hypotheses: B = FATF (Information on Law), C = Consequences (Penalties), D = Appropriateness (Norms)

H4CD(+)

H5D (-)

H6D (-)

Figure 2 Host country conditions model.


606


across experimental conditions. This suggests thatthe randomization worked as expected.

ImplementationThe study was conducted exclusively throughemail, with the target firms receiving either theplacebo or one of the treatments. In consultationwith industry experts, we established that email isthe primary mode of communication for CSPs,which had the benefit of allowing us to reach amuch larger sample (providers in 176 countries)than we would have with other forms of commu-nication. We wrote and sent the emails exclusivelyin English because English is the global language ofbusiness, individuals from our set of ‘‘Norstralia’’countries are assumed to be either fluent or veryproficient in English, and incorporation services aretypically small operations without the capacity towork in a diverse set of different languages. Theseconsiderations all default to the global businesslanguage of English.

We identified the proper content and tone forthe emails to minimize the risk of suggesting thatthis was a scam letter. Ensuring the plausibility ofour email approaches in this manner involveddozens of onsite interviews with CSPs to establishtypical client profiles, as well as attending andobserving CSP trade conferences in New York,Miami, London, Hong Kong, Singapore, Switzer-land, and the Caribbean. To minimize the risk ofdetection, we drafted 33 separate emails, each ofwhich included a salutation, noted brief back-ground information, signaled the intent foranonymous incorporation, and asked the servicewhich materials were necessary to move forward.These four components were included in all of theemails and any other material in the email textwas designed to be purely innocuous. While manyexperiments hold constant all text except thetreatment material, identical language would havesubstantially raised the risk of detection if firms inthe industry forwarded our emails. Within theseemails, we then embedded treatment language,which was unique to each treatment, for each ofthe three conditions specifically noting our aware-ness of international standards, invoking legalpenalties, or raising appropriate behavior (Ap-pendix D provides four examples of emails withspecific treatment language highlighted in bold).Thus we used different email bodies and insertedthe standard treatment language for each condi-tion, and then in post-estimation tested for letter-specific fixed effects. While a few of the letter texts

indicated significant fixed effects (as would beexpected by random chance with so many signif-icance tests), controlling for the imbalances didnot qualitatively change the effects for the exper-imental conditions. All emails were sent from a‘‘consultant’’ indicating a desire to incorporate afirm in pursuit of lower tax obligations, withlimited legal liability, and heightened confiden-tiality. We then coded the replies to our requeststo determine levels of compliance.

Coding ProtocolFull compliance with FATF standards requires thatin order to incorporate, certified identity docu-ments, including at least one notarized photodocument, be received and verified, and that thisdocumentation be retained so that true ownershipcan be determined should the need emerge in thefuture. The lack of such documentation means thatthe actual owners or those who control the com-pany cannot be identified, which effectively createsan anonymous corporation that can be used forillegitimate purposes.Email responses were coded as either Compliant,

Partially Compliant, or Non-Compliant with interna-tional standards, or as Refusal of Service. Responseswere deemed Compliant if the incorporation firmrequired photo identification that was notarized,certified, or apostilled. Partially Compliant firmsrequired photo identification, but did not requireit was notarized or certified. Non-Compliant firmsdid not request any form of photo identity docu-mentation. Incorporation firms that did respondbut refused to offer their services were deemedRefusal of Service. Firms from which we did notreceive a reply were coded as No Response.After the incorporation firms responded with the

necessary information, the ‘‘consultant’’ repliedthat their ‘‘needs have been met’’ and that theprovider’s services would no longer be necessary,effectively terminating the relationship. Althoughwe ended the communication after this correspon-dence, we are confident that the results would nothave changed had we continued through to officialincorporation. In an earlier audit study, one of theauthors followed up with 45 firms all the way to thefinal step and even incorporated (in his own name)three of the firms in Nevada, Britain, and theSeychelles (Sharman, 2011). In no case did theservice providers respond in one way initially andthen change their minds in the end before fullycompleting the transaction. This impression isstrengthened by the extensive interviews with


607


CSPs, who explained that it would be commerciallycounter-productive to send mixed messages toclients as to what documents are required. Fromthese firms’ point of view, they want customers toprovide all the necessary documentation as soon aspossible to complete the transaction quickly, min-imizing the time spent completing customer duediligence.

To prevent subjects from determining that theemails actually originated from the US, researchersused proxy servers that assigned random IPaddresses. All identifying information regardingthe CSPs was removed from emails to ensure theiranonymity and prevent harm to subjects. Uponcompletion of the study, all identifying informa-tion was separated from the results and removed soas to fully cover any trail connecting subjects to theresults.

ANALYSIS AND RESULTSWe first consider the overall results of the auditstudy. In all, 1070 of 2280 emails (or 46.9%) did notreceive responses. For firms that responded, 203replies (or 8.9%) were Non-Compliant, 378responses (or 16.6%) were Partially Compliant,399 replies (or 17.5%) were fully Compliant, and230 responses (or 10.1%) refused to offer incorpo-ration services. The audit results reveal that non-compliance with international standards is aston-ishingly high. If one instead calculates the percent-age Non-Compliant or Partially Compliant, theresults show that 25.5% of firms were not fullyCompliant. When calculated as a percentage ofthose responding, a full 48% were not whollycompliant with international standards.

To understand how these results are distributedacross different country types and therefore distinctinstitutional contexts, we computed levels of non-and partial compliance for the three country cate-gories. When restricting the sample to firms inOECD countries only, 27.4% were not fully com-pliant and of those responding 51.2% were notfully compliant. In tax haven countries, 23.6% werenot fully compliant and of those responding 35.5%were not fully compliant. In developing countries,24.7% are not fully compliant and of thoseresponding 55.6% were not fully compliant. Thesedata show that tax havens have a significantlylower proportion of firms that are out of fullcompliance with international standards, whereasOECD and developing countries are more similar incompliance levels.

To test our experimental hypotheses using thedata from the field experiment, we conducted acomparison of means (t tests) of the three treat-ment conditions against the placebo. We thenstratified the larger sample into the three countrycategories (OECD, Tax Haven, and Developing) andcompleted a similar comparison of means withinthe country blocks. Although the test of thehypotheses could generally use a one-tailed t test,we took the more conservative approach and use atwo-tailed test for all our analyses. See Table 1 forthe data and analyses.While the hypotheses specifically focus on the

treatment effects for compliance with the law, theycould be expanded to also address the relationshipwith the other responses. For example, since notreplying might be viewed as a form of soft compli-ance, one might expect a similar relationship withno reply as full compliance. Refusal, while notprecisely compliant with the law, indicates anunwillingness to conduct business with a poten-tially risky customer and has the same end result ascompliance: no untraceable shell company is pro-vided. As noted below, although we are unable toconclusively ascertain the reason for not replying,an outright refusal is easier to interpret.

Soft ComplianceFirms that did not reply to our initial email requestor to one of the two follow-up communications,and did not have the email returned undeliverable,could be deemed as having been in soft compliance.Since we do not have any details on the reasons fortheir not replying, we are unable to conduct morecomplete analysis of the motives for such softcompliance. However, we did perform a non-response check on all firms that failed to reply toany of our repeated inquiries. The inquiries basi-cally asked whether the firm was still in businessand assisting international customers and made nomention of taxes, legal liability, or confidentiality.A very small proportion of non-responding firmsreplied to this inquiry (less than 5%), suggestingthat soft compliance may not in fact account formuch of response rate beyond the variance acrossexperimental conditions.For many of the target firms, we assume that the

potentially dubious nature of the solicitation,regardless of the placebo or treatment condition,prompts some firms to simply ignore our requestand not provide an explanation for why, leading tosoft compliance. Out of the overall sample, 1070 of2280 requests, or 46.9%, did not respond to our


608


email. For the total sample, there is no significantdifference between the three treatments and theplacebo for no reply, with all four conditions fallingwithin 2.6% of the average across all conditions.For the three country conditions, there is greatervariance for the placebo no reply condition (44.7,34.8, and 51.1% for OECD, Tax Haven, and Devel-oping countries, respectively), and these differencesare significant statistically at the p\0.05 level.

Within the country subgroups, for OECD coun-tries, there is no significant difference between theplacebo and the three treatments for no reply. Fordeveloping countries, the FATF treatment (64.6

versus 51.1%, p\0.01) and the appropriatenesstreatment (61.5 versus 51.1%, p\0.05) are signif-icantly higher than the placebo for no reply. For TaxHaven countries, we see significant differences inthe non-response rate for all three conditions, withthe FATF treatment having 10% less no reply (24.7versus 34.8%, p\0.10), the consequences treatment9.5% higher (p\0.10) and appropriateness 11.9%lower (p\0.05) than the placebo condition. Aninteresting take away from these findings showsthat in Tax Havens significantly more firms reply(the opposite outcome from no reply) wheninformed of international law or norms and less

Table 1 Displays difference in means tests for the total sample, OECD countries, Tax Haven countries, and Developing countries (row

blocks) for the five possible outcomes of no-reply, non-compliant, partial compliant, compliant, and refusal (rows within blocks) for

the Placebo, FATF, Consequences, and Appropriateness experimental conditions (columns)

Response Placebo FAFT (information on law) Sig. Rationalism Sig. Contructivism Sig.

(Penalties) (Norms)

Comparison of means between placebo and treatments - total sample

No reply 44.75% 49.46% 48.63% 50.00% +

Non-compliant 8.42% 8.11% 7.38% 9.78%

Partial compliant 16.65% 16.22% 19.95% 14.13%

Compliant 19.49% 16.76% 16.39% 16.03%

Refusal 10.69% 9.46% 7.65% * 10.05%

Total 100.00% 100.00% 100.00% 100.00%

n 1057 370 366 368

Comparison of means between placebo and treatments – OECD countries

No reply 44.72% 46.81% 45.06% 52.25%

Non-compliant 13.73% 14.89% 10.99% 11.71%

Partial compliant 15.14% 13.83% 15.39% 11.71%

Compliant 14.09% 8.51% 18.68% 10.81%

Refusal 12.32% 15.96% 9.89% 13.51%

Total 100.00% 100.00% 100.00% 100.00%

n 284 94 91 111

Comparison of means between placebo and treatments - Tax Haven countries

No Reply 34.78% 25.74% + 45.28% + 22.89% *

Non-Compliant 4.01% 4.95% 3.77% 8.43%

Partial Compliant 15.72% 24.75% * 22.64% 18.07%

Compliant 39.80% 40.59% 23.59% ** 36.15%

Refusal 5.69% 3.96% 4.72% 14.46% **

Total 100.00% 100.00% 100.00% 100.00%

n 299 101 106 83

Comparison of means between placebo and treatments - developing countries

No reply 51.06% 64.57% ** 52.66% 61.49% *

Non-compliant 8.02% .29% .69% .20%

Partial compliant 18.14% 2.57% + 0.71% 3.79%

Compliant .92% 7.43% 0.65% .77%

Refusal 12.87% 9.14% 8.28% 5.75% *

Total 100.00% 100.00% 100.00% 100.00%

n 474 175 169 174

Notes: Each entry refers to the percent of subjects in a country block who received a given condition and responded in a certain way. Those percentagesare accompanied by significance tests, which compare the FATF, Consequences, and Appropriateness conditions to the Placebo and identify when theyare meaningfully different outcomes.

Sig. significance of difference between mean of Placebo and each treatment.+ p\0.10; * p\0.05; ** p\0.01; *** p\0.001.


609


when informed of penalties than the placebocondition.

FATF (Information on Law) Treatment(Hypothesis 1)For the FATF treatment, we primed the incorpora-tion firm with the information that identified FATFand the pertinent international standards on cor-porate transparency. Hypothesis 1 states that wheninformation about international standards andlaws is provided, the CSP is more likely to complywith them than when no information is provided.As seen in Table 1, there are no significant differ-ences between the FATF treatment and the placebofor the full sample of responses, including fullcompliance, failing to support H1. For this treat-ment, no reply increases from 44.5% in the placeboto 49.0%, but compliant drops from 18.9% to 16.9%and refusal (a different sort of compliance) alsodeclines from 11.2% to 9.5%, but none of thesechanges are significant statistically.

Consequences (Penalties) Treatment (Hypothesis2)For the Consequences treatment, we providedsimilar details with the FATF treatment, but addedinformation about potential financial penaltiesthat could result from non-compliance. Hypothesis2 predicts higher compliance will result whenconsequences for violation to international stan-dards are raised. Only refusal had a significantchange (a decrease to 7.6%, p\0.05), but this is inthe opposite direction as expected, failing to sup-port this hypothesis (H2). As with the InternationalLaw treatment, no reply increases (to 49.2%) andCompliant drops (to 16.0%), but neither are signif-icant changes. These decreases in Compliant andRefusal seem to be offset by the increase in the noreply (soft compliance) and Part Compliant.

Appropriateness (Norms) Treatment(Hypothesis 3)For the Appropriateness treatment, we appealed tothe target’s sense of doing what is right by suggest-ing that compliance with international law bebased on norms of appropriateness and generalconformity. The Norms Hypothesis asserts highercompliance when norms of appropriateness areinvoked, but as with the other two treatments, wefind compliance actually decreases (to 15.5%), butinsignificantly, failing to support H3. We do find amarginal significantly higher rate for no reply,which increased to 49.5% (p\0.10). One

explanation may be that ‘‘accepted norms’’ maybe not be universally agreed upon, and as withculture, may differ across countries (Leung &Morris, 2015).These results provide no support for any of the

primary treatment hypotheses and only find asignificant relationship for refusal responses underthe consequences treatment, but in the oppositedirection as expected. Greater insight into whythese relationships are not found may be gained bystratifying the total sample into country categoriesto determine if the nature of the institutionalcontext in the country in which the CSP is basedaffects behavior differently. As noted above, thefield experiment was deliberately designed in theuse of block randomization so that we could test iffirms in OECD, Tax Haven, and Developing coun-tries responded differently to our treatments.

Tax Haven Countries (Hypothesis 4)We hypothesized that CSPs in tax haven countriesare less likely to violate international law and havehigher compliance than firms in OECD and devel-oping countries (H4A). We also hypothesized that,when informed of international standards (H4B),informed of the consequences for violating theseinternational standards (H4C), and when thenorms of appropriate behavior to follow interna-tional standards are appealed to (H4D), tax havenswould be more sensitive to these prompts priminginternational law. The overall soft compliance rates(no reply) were significantly lower in tax havensthan OECD and Developing countries (33.5 versus46.6 and 55.5%, respectively). This means that CSPsin tax havens were more likely to respond to ourrequests than in OECD countries and developingcountries. Nearly two-thirds of tax-haven firmsresponded, in contrast to less than half of develop-ing country firms. It seems like the increasedresponse rate is found in a significantly highercompliance rate (36.5% for tax havens versus13.3% for OECD countries and 9.6% for developingcountries). Firms in tax havens are nearly threetimes more likely to be fully compliant than firmsin OECD countries and nearly four times morelikely than firms in developing countries. AlthoughCSPs in tax havens have lower refusal rates than inOECD and developing countries, this seems a resultof much higher compliance rates, supporting H4A.Within the tax haven countries, we find that for

the International Law treatment, the no reply rate ismarginally significantly lower (p\0.10) and thepartial compliant rate significantly higher (p\0.05)


610


than the placebo condition, but we do not findmuch of a change in full compliance. In contrast, forthe consequences treatment, the no reply rate margin-ally significantly increases (p\0.10) and the com-pliant rate significantly drops (p\0.01). For theappropriateness treatment, no reply is also signifi-cantly lower (p\0.05) and the refusal rate issignificantly higher (p\0.01). When informed ofinternational law, tax-haven firms respond to ourrequest more often and have increasing partial, butnot full compliance. When informed of penalties,hard compliance drops, but this seems to bebecause of a shift to the soft compliance of notresponding to our request. When accepted normsare invoked, more firms respond, but then explic-itly refuse to do business with us. Interestingly,while just out of marginal significance (p\0.11),the penalties condition encourages higher partialcompliance and the norms condition results inhigher non-compliance. These results provide somesupport for Hypothesis 4 considering CSPs in taxhavens, which demonstrated significantly highercompliance with global standards and more sensi-tivity to prompts about those standards than firmsin OECD and developing countries.

OECD Countries (Hypothesis 5)Hypothesis 5 predicted that, compared to taxhavens, CSPs based in OECD countries would beless likely to comply with international standards(H5A) and would prove less sensitive wheninformed of international standards (H5B),informed of the consequences for violating theseinternational standards (H5C), and when thenorms of appropriate behavior to follow interna-tional standards are appealed to (H5D). While thereis some variance seen across the treatments whenconsidering response rates, there are no significantdifferences between the placebo condition and anyof the treatments for all five outcome measures,offering only support for Hypothesis 5A.

Developing Countries (Hypothesis 6)As discussed above, developing countries face dif-ferent pressures than OECD and tax-haven coun-tries. While they are pressured to accept the sameglobal norms for laws and transparency, they mayneither have the resources to enforce the laws, norbe dissuaded by potential penalties. Thus theDeveloping Country Hypothesis proposed that,compared to firms in tax havens, incorporationfirms in developing countries should be less likelyto comply with international standards (H6A) and

be less sensitive when informed of internationalstandards (H6B), when notified of the conse-quences for violating these international standards(H6C), and when the request appeals to norms ofappropriate behavior in following internationalstandards (H6D). While we find support for H6A,we find that none of the conditions create asignificant change in full compliance over theplacebo. We do find that when informed of inter-national law, developing country firms are lesslikely to even respond, with no reply increasingfrom 51.1 to 64.6% (p\0.01), and are less likely tobe partially compliant, decreasing from 18.1 to12.6% (p\0.10). The penalties condition resultsin no significant changes. The norms conditionalso exhibited a significant increase in no reply (to61.5%, p\0.05) and a significant drop in refusalfrom 12.9 to 5.8% (p\0.05). Firms in developingcountries appear more sensitive to the primingconditions than firms in OECD countries, but thesensitivity we do see appears to cut in oppositedirections from firms in tax havens.The results of these comparison-in-means tests

are robust qualitatively to logistic regression anal-ysis, considering each outcome category separatelywith the other outcomes as the comparison sets.Additionally, although it has not yet becomestandard practice in social science or businessstudies, statistical best methods recommend theemployment of corrections for multiple compar-isons given that significant results at the p\0.05level should be expected in one of twenty testsbased on chance alone. We therefore made use oftwo methods to adjust for multiple comparisons.We employed the more conservative family-wise-error-rate adjustment using the Bonferroni method(Dunn, 1961), and we also controlled for the false-discovery rate with the less draconian Benjamini–Hochberg procedure (Benjamini & Hochberg,1995). As is common, many of the results did notclear the amended critical p-value thresholds forsignificance, but other key results held. Notably, forthe tax haven subgroup, the effects were robust toboth multiple testing corrections of the normstreatment in significantly increasing refusal ratesand the penalties condition in significantly decreas-ing the compliance rate (presumably because thefirms that would otherwise have proved compliantdid not reply).6

All told, the subgroup analysis provides supportfor the importance of institutional context, partic-ularly regarding tax havens. Firms in tax havenswere significantly more compliant than firms in


611


OECD or developing countries and significantlymore sensitive to prompts regarding internationallaw, formal penalties for violating the standards,and informal norms enjoining conformity. Theseresults underscore the need for studies that aresensitive to institutional context but also for morenuanced categorization of countries according totheir rule of law and position in the global econ-omy, at least considering tax havens as a distinctcategory when studying international finance,incorporation, and law.

DISCUSSION AND CONCLUSIONSThe audit study and accompanying field experi-ment offer interesting results and important impli-cations for scholars and national policymakers. Wefirst considered overall levels of firms’ complianceand found that they neither match well the FATFnational level assessments, nor do they offer aparticularly optimistic view about financial andlegal institutions in global corporate governance.This key finding from the audit study extendsbeyond our earlier work (Findley et al., 2014),which was almost exclusively experimental andfocused on different experimental treatments,specifically ones related to terrorism, corruption,the IRS, and offering a premium. In particular itextends the work to a careful consideration of theimportance of institutional context in conditioningcompliance with global corporate transparencystandards.

As shown by the 2016 Panama Papers scandaland this article, it seems that the efforts theinternational community is putting into improvingcorporate transparency are not generally working.It also appears that it is not very difficult to createan anonymous shell corporation. While it is ini-tially encouraging that nearly 75% of the totalsample of incorporation firms either did not reply,proved fully compliant, or refused service, that stillleaves roughly 25% of these firms that do notcomply or are only partially compliant. The infer-ence is that, should potential criminals wish to setup an untraceable shell company, all they need todo is shop around and they would succeed, onaverage, within four attempts.

We then divided the sample to learn whethernational characteristics help explain behavior, con-sistent with institutional arguments in interna-tional business. We found compelling evidence forinstitutional differences across country types. ForOECD countries, the lack of significant changes in

any responses across the three treatments wasunexpected and concerning. These countries’claims to be leading the way in adhering to theglobal rules they created to counter money laun-dering and associated crimes does not find evidencein the data. The overall full compliance rate forOECD countries is 13.8%, which is higher than the9.6% for developing countries, but is significantlylower than Tax Havens (36.5%). With 27.4% ofincorporation firms either being non-compliant oronly partially compliant, it apparently does nottake much effort to set up an anonymous shellcorporation in an OECD country.In contrast, firms in tax havens, where we

initially expected greater challenges to setting upan untraceable shell corporation, had actual com-pliance rates of more than a third, and had thelowest non-compliance rate of any category (4.8versus 13.1% for OECD countries and 7.9% fordeveloping countries). Incorporation firms in TaxHavens are thus found to be more likely to complywith international law for incorporation. Oneexplanation may be that small, vulnerable taxhaven countries and firms are responding to theincreased scrutiny from the rest of the world byimproving corporate governance and transparency(Sharman, 2010, 2011).For the experimental portion of the study, we

hypothesized that CSPs, when given informationabout international law, potential penalties, andgenerally accepted norms, would be more likely tocomply with the law. We found that this is not thecase on the whole. In fact, when primed aboutinternational law, penalties, and norms, compli-ance to the law and refusal to proceed with ourrequest actually decreased, something we did notexpect. We did find that in all three treatments, theincorporation firm’s soft compliance, as measuredby the rate of non-response, did increase, but onlywith marginal significance for the appropriatenesstreatment. It seems that these treatments did notsignificantly increase the likelihood for hard com-pliance to international laws across all countrygroups.However, for the treatment effects in tax havens,

the no reply was marginally to highly significant forall three treatments, with a decrease in no reply forthe FATF and appropriateness treatments, but anincrease for the consequences treatment. For thelatter, it appears that the increase in non-responsecame out of a decrease in compliance. One defen-sible inference is that otherwise compliant firms intax havens, when informed of penalties, are


612


basically ignoring our requests and not responding.In contrast, when primed with acceptable norms,tax haven firms are more likely to respond, but alsomore likely to refuse service outright, perhaps inresponse to the apparent disingenuousness of theappeal to appropriateness.

Although the tax havens provide the mostintriguing results and those most consistent withinstitutional theories of international business,incorporation firms in developing countries alsoprovide some interesting insights. Developingcountries had the highest overall no reply rate(55.5%), and within this category, the internationallaw and norms treatments had even significantlyhigher non-response rates over the placebo. Overallcompliance rates were low (9.6%) in developingcountries and the treatments did not significantlyimprove compliance but rather, in multiple cases,worsened adherence to global standards. Thismight be explained as resulting from these coun-tries’ lack of resources or incentives to enforceinternational law.

One key implication of our findings is that OECDcountries need to do more to get their own housesin order. Our study argues that the low overall rateof compliance and the practical ease of obtainingprohibited shell companies in OECD countries doesnot call for new laws, but for re-doubled efforts tomake these laws effective.

LIMITATIONS AND FUTURE RESEARCHAlthough the use of an audit study with embeddedfield experiment helps overcome many concernsthat might be leveled at other methodologies thatmight be used to study this phenomenon, such assurveys or lab studies, other limitations for thisstudy should be identified. As noted above, oursample selection relied heavily on Internet searchesto compile a subject pool of CSPs around the world,since no database for such firms is available. Werecognize that we were unable to document thecomplete population of CSPs, but are comfort-able with the sample we ultimately developed. Wewere able to create a large sample of firms involvedwith providing incorporation services with broadcoverage in terms of countries and regions. Theblocking and randomization process did not favorany type of firm, country, or region, allowing highconfidence in and strong generalization from theresults. In addition, since the firms we were able toidentify are more visible, we believe our sample

provides for a more conservative test, since wesuspect that firms that may intentionally be diffi-cult to find are more likely to violate internationallaw.Due to the nature of our field study, we were

unable to further explore or better understand thelogic for why a firm did not reply. There may havebeen an element of soft compliance, but a follow-up round of communication did not offer strongevidence of this. For the tax haven countries, the noreply rate varied significantly, from as low as 22.9%for appropriateness to as high as 45.3% for conse-quences, so we are confident that the treatments didaffect the firm’s propensity to reply and whichresponse they provided. Such variance for no replyhad implications on other responses, and gaining abetter grasp for why a firm did not reply would behelpful to further understand the treatment effectsin light of selection into response.We recognize that the client profiles we use in

this study are from developed countries, limitingthe conclusions we can draw accordingly. Thisdesign was intentional to ensure consistency fromthe incorporating firm perspective. While it isexpected that there is a wide variety of potentialcustomers wishing to set up shell companies, insetting up our customer profiles we have aimed toreplicate a reasonably typical shell-company buyer.In particular, one interested in limited liability, taxminimization, and confidentiality/secrecy from adeveloped country. Future research could furtherexplore the implications from having the incorpo-rating entity based in a wider range of countries,including developing countries.While we adopted the conventional country

classifications of OECD, Tax Haven, and Developingcountries to segment our sample and test ourhypotheses, there are other categories that wouldalso be interesting to consider. Future research couldsee if the standard of law, whether civil or common,has an influence on compliance or results in differ-ent subgroup effects for the treatments. Futureresearch could also consider whether there is a pricepremium associated with violating internationalincorporation standards and whether the treatmentscause change to the cost of incorporation.Although email is the standard mode of commu-

nication for this type of transaction, phone or in-person contact occurs to a lesser extent. Inferencesfrom our approach, while capturing the dominantmode of communication, should not be made fornon-email correspondence.


613


Finally, what became of Vadim Mikerin? He wasarrested on October 29, 2014 and was charged withsoliciting bribes from lobbyists. He agreed to pleadguilty to 20 charges of money laundering andbribery. These bribes and other payments, esti-mated to be worth $1.7 million, were allegedlyfunneled through shell corporations; shell corpo-rations that could have been set up anonymouslyin as few as four attempts.

ACKNOWLEDGEMENTSThe authors would like to thank Brian Silverman, SteveTallman, Douglas Cumming, and three anonymousreviewers for their insightful comments. An earlierversion of this article was presented at the Journal ofInternational Business Studies’ Paper DevelopmentConference in London, UK (February 2016).

NOTES

1If the latter purposes are intended, anonymity andlack of transparency is desired and generally necessaryto succeed. Here is where the challenge for nationaland international policymakers comes into play. Tohighlight this challenge, a group of United StatesSenators issued a report in 2013 on ways to stop thelaundering of ‘‘blood money’’ from drug traffickers(United States Senate Caucus on International Nar-cotics Control, 2013). One of the key recommenda-tions was to make it more difficult for criminal groupsto set up and use shell corporations to launder theirmoney in the US and throughout the global financialsystem. The report concludes that ‘‘it is far too easy forUS corporations to have ‘hidden owners’’’ (p. 24) andnotes concerns that an anonymous corporation can beset up using less information than is required to open abank account or get a driver’s license. Nationalgovernments want the domestic corporate entitiesthat have been granted legal status, rights andprivileges to follow clear rules in their creation andsubsequent behavior. They should be ongoing con-cerns with real business activities and there should alsobe a clear and certified knowledge of who created theentity, who the current corporate officers are, and howto contact them. Laws are enacted, and hopefullyenforced, to ensure these things are accomplished forthe establishment and monitoring of corporations. At

the international level, standards have been created sothat general consistency across country law andpractice is encouraged and maintained. This allowsfor greater reliability within and across countries andcan help address the added concerns and loss ofcontrol that may come about when dealing with crossborder activities, particularly illicit ones.

2In April 2016, a leak of 11.5 million documentsregarding over 200,000 offshore accounts from thePanamanian corporate service provider and law firmMossack Fonseca revealed the widespread use ofshell companies for tax avoidance, evasion of inter-national sanctions, money laundering, and fraud.Several heads of state and their associates, as well aswealthy individuals were implicated in the docu-ments. While many were not actually engaged inillegal activities, their use of shell companies toshelter or hide certain activities and funds came asan embarrassment and exposed the pervasive use ofshell corporations.

3See Appendix A for definition of key terms.4Private shell companies represent the vast majority

of shell companies. To highlight this point, there arenearly 6000 companies collectively listed on the NYSEand NASDAQ stock exchanges, but more than800,000 companies incorporated in the British VirginIslands alone. Firms that list publicly have to meetexacting disclosure and reporting requirement. Privateshell companies, the target of recent global regulatoryinitiatives and the focus of our article, are not. Fortune500 companies and their smaller counterparts areengaged in producing substantive goods and services,whereas the large majority of shell companies do not,hence the ‘‘shell’’ moniker.

5The FATF standards were slightly modified inFebruary 2012 and the various recommendations re-numbered. The substance of both these internationalstandards and the domestic laws of key jurisdictionsremained constant across the period of the auditstudy, MERS, and field experiment. Post-2012, Rec-ommendations 5, 12, and 33 became 10, 22 and 24,respectively.

6For developing countries, the significant decreasein refusal rates for the norms condition survived theBenjamini–Hochberg correction but not the Bonfer-roni adjustment, as did the increase in no-reply rates inthe norms condition. The increase in non-response tothe FATF condition in developing countries provedrobust to both multiple-testing corrections.


614


REFERENCESAbbott, K. W., Keohane, R. O., Moravcsik, A., Slaughter, A., &Snidal, D. (2000). The concept of legalization. InternationalOrganization, 54(3): 401–419.

Allred, B. B., & Park, W. G. 2007. Patent rights and innovativeactivity: Evidence from national and firm-level data. Journal ofInternational Business Studies, 38: 878–900.

Arend, A. C. 1997. Do legal rules matter? International law andinternational politics. Virginia Journal of International Law, 38:107–154.

Bandiera, O., Barankay, I., & Rasul, I. 2011. Field experimentswith firms. Journal of Economic Perspectives, 25(3): 63–82.

Banerjee, A. V., & Duflo, E. 2009. The experimental approach todevelopment economics. Annual Review of Economics, 1:151–178.

Barney, J. B. 1991. Firm resources and sustained competitiveadvantage. Journal of Management, 17(1): 99–121.

Benjamini, Y., & Hochberg, Y. 1995. Controlling the falsediscovery rate: A practical and powerful approach to multipletesting. Journal of the Royal Statistical Society, 57(1): 289–300.

Bertrand, M., & Mullainathan, S. 2004. Are Emily and Gregmore employable than Lakisha and Jamal? A field experimenton labor market discrimination. The American Economic Review,94(4): 991–1013.

Birkinshaw, J., Bresman, H., & Nobel, R. 2010. Knowledgetransfer in international acquisitions: a retrospective. Journal ofInternational Business Studies, 41: 21–26.

Buckley, P., & Casson, M. 1976. The future of the multinationalenterprise. London: Palgrave Macmillan.

Butler, D. M., & Broockman, D. E. 2011. Do politicians raciallydiscriminate against constituents? A field experiment on statelegislators. American Journal of Political Science, 55(3): 463–477.

Cantwell, J., Dunning, J. H., & Lundan, S. M. 2010. Anevolutionary approach to understanding international busi-ness activity: The co-evolution of MNEs and the institutionalenvironment. Journal of International Business Studies, 41:567–586.

Chatterji, A. K., Findley, M., Jensen, N. M., Meier, S., & Nielson,D. 2016. Field experiments in strategy research. StrategicManagement Journal, 37(1): 116–132.

Chayes, A., & Chayes, A. H. 1996. The new sovereignty:Compliance with international regulatory arrangements. Cam-bridge, MA: Harvard University Press.

Checkel, J. T. 2001. Why comply? Social learning and Europeanidentity change. International Organization, 55(3): 553–588.

Chilton, A., & Tingley, D. 2013. Why the study of internationallaw needs experiments. Columbia Journal of Transnational Law,52: 173–237.

Coffee, J. C., Jr. 2001. Do norms matter? A cross-country evalua-tion. University of Pennsylvania Law Review, 149(6): 2151–2177.

Demirguc-Kunt, A., Love, I., & Maksimovic, V. 2006. Businessenvironment and the incorporation decision. Journal of Bank-ing & Finance, 30(11): 2967–2993.

Djelic, M., & Quack, S. 2003. Globalization and institutions:Redefining the rules of the economic game. Northampton, MA:Edward Elgar.

Drezner, D. W. 2007. All politics in global: Explaining internationalregulatory regimes. Princeton, NJ: Princeton University Press.

Duflo, E., Glennerster, R., & Kremer, M. 2008. Using random-ization in development economics research: A toolkit. In T.P. Schultz & J. Strauss (Eds.), Handbook of DevelopmentEconomics (Vol. 4, pp. 3895–3962). Amsterdam: Elsevier.

Dunn, O. J. 1961. Multiple comparisons among means. Journalof the American Statistical Association, 56(293): 52–64.

FATF. 2013. International standards on combating moneylaundering and the financing of terrorism & proliferation.www.fatf-gafi.org/media/fatf/documents/recommendations/pdfs/FATF_Recommendations.pdf, accessed 29 October 2016.

FATF. 2016. www.fatf-gafi.org/pages/aboutus/, accessed 16September 2016.

Findley, M. G., Nielson, D. L., & Sharman, J. C. 2014. Globalshell games: Experiments in transnational relations, crime, andterrorism (Vol. 128). Cambridge: Cambridge University Press.

Findley, M. G., Nielson, D. L., & Sharman, J. C. 2015. Causes ofnon-compliance with international law: A field experiment ofanonymous incorporation. American Journal of Political Science,59(1): 146–161.

Galang, R. M. N. 2012. Government efficiency and internationaltechnology adoption: The spread of electronic ticketingamong airlines. Journal of International Business Studies, 43:631–654.

Gerber, A. S., & Green, D. P. 2012. Field experiments: Design,analysis and interpretation. New York: W.W. Norton.

Globerman, S., & Shapiro, D. 2003. Governance infrastructureand US foreign direct investment. Journal of InternationalBusiness Studies, 34: 19–39.

Goldsmith, J. L., & Posner, E. A. 2005. The limits of internationallaw (Vol. 199). Oxford: Oxford University Press.

Goldstein, J., Kahler, M., Keohane, R. O., & Slaughter, A. 2000.Introduction: Legalization and world politics. InternationalOrganization, 54(3): 385–399.

Granovetter, M. 1985. Economic action and social structure:The problems of embeddedness. American Journal of Sociology,91(3): 481–510.

Harrison, G. W., & List, J. A. 2004. Field experiments. Journal ofEconomics Literature, 42(4): 1009–1055.

Hiller, D., Pindado, J., de Queiroz, V., & de la Torre, C. 2011.The impact of country-level corporate governance on researchand development. Journal of International Business Studies, 42:76–98.

Hoskisson, R. E., Eden, L., Lau, C. M., & Wright, M. 2000.Strategy in emerging economies. Academy of ManagementJournal, 43(3): 249–267.

Hyde, S. D. 2007. The observer effect in international politics:Evidence from a natural experiment. World Politics, 60(1):37–63.

Ingram, P., & Silverman, B. 2002. Introduction. In P. Ingram &B. Silverman (Eds.), The new institutionalism in strategicmanagement (pp. 1–30). Amsterdam: Elsevier.

Kostova, T., Roth, K., & Dacin, M. T. 2008. Institutional theory inthe study of multinational corporations: A critique and newdirections. Academy of Management Review, 33(4): 994–1006.

Leung, K., Bhagat, R. S., Buchan, N. R., Erez, M., & Gibson, C. B.2005. Culture and international business: Recent advancesand their implications for future research. Journal of Interna-tional Business Studies, 36: 357–378.

Leung, K., & Morris, M. W. 2015. Values, schemas, and norms inthe culture–behavior nexus: As situated dynamics framework.Journal of International Business Studies, 46: 1028–1050.

Levitt, S. D., & List, J. A. 2009. Field experiments in economics:The past, the present, and the future. European EconomicReview, 53(1): 1–18.

March, J. G., & Olsen, J. P. 1998. The institutional dynamics ofinternational political order. International Organization, 52(4):943–969.

Meyer, K. E., Estrin, S., Bhaumik, S., & Peng, M. W. 2009.Institutions, resources, and entry strategies in emergingeconomies. Strategic Management Journal, 30(1): 61–80.

Meyer, K. E., & Peng, M. E. 2005. Probing theoretically intoCentral and Eastern Europe: Transactions, resources, andinstitutions. Journal of International Business Studies, 36:600–621.

Meyer, K. E., & Sinani, E. 2009. When and where does foreigndirect investment generate positive spillovers? A meta-analy-sis. Journal of International Business Studies, 40(7): 1075–1094.

Miller, K. 1992. A framework for integrated risk management ininternational business. Journal of International Business Studies,23: 311–331.

North, D. C. 1990. Institutions, institutional change and economicperformance. Cambridge: Cambridge University Press.


615


http://www.fatf-gafi.org/media/fatf/documents/recommendations/pdfs/FATF_Recommendations.pdf

http://www.fatf-gafi.org/media/fatf/documents/recommendations/pdfs/FATF_Recommendations.pdf

http://www.fatf-gafi.org/pages/aboutus/

OECD. 2009. Countering Offshore Tax Evasion. www.oecd.org/ctp/harmful/42469606.pdf, accessed 29 October 2016.

Peng, M. W., Wang, D. Y. L., & Jiang, Y. 2008. An institution-based view of international business strategy: A focus onemerging economies. Journal of International Business Studies,39: 920–936.

Porter, M. E. 1980. Competitive strategy. New York: Free Press.Redding, G. 2005. The thick description and comparison ofsocietal systems of capitalism. Journal of International BusinessStudies, 36: 123–155.

Schectman, J. 2015. Russian uranium sales to US probed. WallStreet Journal, May 1, B1.

Scott, W. R. 1995. Institutions and organizations. ThousandOaks, CA: Sage.

Shadish, W., & Cook, T. 2009. The renaissance of fieldexperimentation in evaluating interventions. Annual Reviewof Psychology, 60: 607–629.

Sharman, J. C. 2010. Shopping for anonymous shell companies:An audit study of financial anonymity and crime. Journal ofEconomic Perspectives, 24: 127–140.

Sharman, J. C. 2011. Testing the global financial transparencyregime. International Studies Quarterly, 54: 981–1001.

Singh, J., Lentz, P., & Nijssen, E. 2011. First- and second-ordereffects of consumers’ institutional logics on firm-consumerrelationships: A cross-market comparative analysis. Journal ofInternational Business Studies, 42: 307–333.

Stephan, U., & Uhlaner, L. M. 2010. Performance-based vssocially supportive culture: A cross-national study of descrip-tive norms and entrepreneurship. Journal of InternationalBusiness Studies, 41(8): 1347–1364.

Sugathan, A., & George, R. 2015. The influence of governanceinfrastructure and corporate governance on profit sharing.Journal of International Business Studies, 46: 886–916.

Teegen, H., Doh, J., & Vachani, S. 2004. The importance ofnongovernmental organizations (NGOs) in global governanceand value creation: An international business research agenda.Journal of International Business Studies, 35: 463–483.

Thompson, A. G. 1996. Compliance with agreements in cross-cultural transactions: Some analytical issues. Journal of Inter-national Business Studies, 27: 375–390.

United States. 1978. The Belmont Report: Ethical principles andguidelines for the protection of human subjects of research.Bethesda, MD: The Commission.

United States Senate Caucus on International Narcotics Control.2013. The buck stops here: Improving US anti-money launderingpractices.

Wang, C., Hong, J., Kafouros, M., & Wright, M. 2012. Exploringthe role of government involvement in outward FDI fromemerging economies. Journal of International Business Studies,43(7): 655–676.

Wernerfelt, B. 1984. A resource-based view of the firm. StrategicManagement Journal, 5(1): 171–180.

Wood, S., Abbott, K. W., Black, J., Eberlein, B., & Meidinger, E.2015. The interactive dynamics of transnational businessgovernance: A challenge for transnational legal theory.Transnational Legal Theory, 6: 333–369.

Wright, M., Filatotchev, I., Hoskisson, R. E., & Peng, M. W.2005. Strategy research in emerging economies: Challengingthe conventional wisdom. Journal of Management Studies,42(1): 1–33.

Anonymous shell corporation A shell corporation for which the beneficial (real) owner has been been disguised, generally in order

to operate without the scrutiny of law enforcement and public. The term shell corporation and shell

company is used interchangeably in the manuscript

Compliant CSPs were deemed compliant if they required photo identification that was notarized/certified to

complete the incorporation

Corporate service provider

(CSP)

A firm or sole proprietor that can incorporate a company on behalf of its founding members and can

also act as the resident agent for the company

Developing countries A country in which the majority lives on far less money – with far fewer basic public services – than the

population in highly industrialized countries

Field experiment A scientific methodology that experimentally examines an intervention in naturally occurring

environments (real world), rather than in the laboratory

Financial action task force

(FATF)

An international organization created by the world’s leading economic powers to counter money

laundering, was created to promote and enforce corporate transparency standards worldwide

Logic of consequences Adherence to laws and standards occurs only to avoid penalties or to achieve some benefit (March &

Olsen, 1998)

Logic of appropriateness Adherence to international law is a function of and response to accepted norms (March & Olsen,

1998)

Mutual evaluation reports

(MERs)

Assessments conducted by FATF to determine whether a country is compliant (C), largely compliant

(LC), partially compliant (PC), and non-compliant (NC) with each of the 40 Recommendations

Managerial School of

International Law

Most countries follow most international rules most of the time (Chayes & Chayes, 1996)

No response Firms from which we did not receive a reply were coded as no response

Non-compliant CSPs were deemed non-compliant if they did not request any photo identity to complete the

incorporation

Norstralian countries Wealthy, low-corruption OECD countries (Norway, Finland, Sweden, Denmark, the Netherlands,

Austria, New Zealand, or Australia). The Norstralian countries are among the least corrupt countries

ranked on the Transparency International Corruption Perceptions Index (CPI) and are where the

consultants in our study claimed to be from

APPENDIX A: DEFINITION OF KEY TERMS


616


http://www.oecd.org/ctp/harmful/42469606.pdf

http://www.oecd.org/ctp/harmful/42469606.pdf

Appendix B provides a summary of the FATFMERs for countries throughout the world, focusingon the three key FATF recommendations. Theyprovide four outcome categories, and thetable shows the distribution of compliance levelsacross the countries and recommendations. Only asmall percentage of countries were Largely Com-pliant and only Recommendation 33 had anycountries identified as fully Compliant. The vastmajority of country audits indicated either Non-Compliance or Partial Compliance with the trans-parency standards.

Appendix C shows the results of regressionanalysis considering whether national level com-pliance with each of the recommendations predictsfirm level compliance. Each of the

‘‘Recommendation’’ variables is an ordered scaleranging from non-compliance to full compliance(1 = non-compliant, 2 = partially compliant,3 = largely compliant, and 4 = compliant) fornational level assessments and predict percentfirm-level compliance. The model is estimatedusing ordinary least squares regression analysis.National compliance does not predict firm-levelcompliance for FATF Recommendations 5 and 12and predicts virtually none of the variance. ForRecommendation 33, national compliance is asignificant predictor (at the p\0.05 level), but lessthan 4% of the variance is explained, offering weaksubstantive support for the proposition that statu-tory law drives firm-level behavior across the fullsample.

OECD countries A collection of 34 countries with the world’s most advanced and developed economies that have committed to

enact and enforce its standards for transparency in corporate governance. These countries adhere to the rule of law

and have relatively well-funded, capable judicial, regulatory and law enforcement

Partially

compliant

CSPs were deemed partially compliant if they requested photo identification to complete the incorporation, but did

not ask the it be notorized/certified

Refusal of

Service

Incorporation firms that did respond, but refused to offer their services were deemed Refusal of Service

Tax Havens Countries identified by the OECD as those that host companies not engaged in substantive business activities, that

apply low or zero tax rates, do not exchange tax information with other governments, and have little transparency

APPENDIX B: SUMMARY OF FAFT MUTUAL EVALUATION REPORTS AUDITS

Recommendation 10 Recommendation 22 Recommendation 24

Frequency Percentage Frequency Percentage Frequency Percentage

Non-compliant 72 46.2% 114 70.4% 49 30.6%

Partially compliant 71 45.5% 44 27.2% 85 53.1%

Largely compliant 13 8.3% 4 2.5% 20 12.5%

Compliant 0 0.0% 0 0.0% 6 3.8%

Total 156 100.0% 162 100.0% 160 100.0%

Model 1 Model 2 Model 3

Constant 0.579 (0.070) 0.495 (0.066) 0.385 (0.66)

Recommendation 10 -0.039 (0.039)

Recommendation 22 0.014 (0.046)

Recommendation 24 0.063 (0.030)

R2 0.0010 0.0009 0.0399

Adj. R2 -0.0001 -0.0087 0.031

F 0.990 0.090 2.060*

N 102 106 104

+ p\0.10; * p\0.05; ** p\0.01; *** p\0.001

APPENDIX C: REGRESSION ANALYSIS


617


APPENDIX D: EXAMPLE EMAILS USEDIN STUDY

Control (Placebo) Email

Dear [name/company]

I am contacting you as I would like to form aninternational corporation for my consulting firm. Iam a resident of [Norstralia] and have been doingsome international consulting for various compa-nies. We are now growing to a size that makeincorporation seem like a wise option. A lot of ournewer business is in your region.

My two associates and I are accustomed to paying[Norstralia] income taxes, but the rising tax ratesmake incorporation in another country a moreeconomical alternative. Also, our contracts growlarger and more complicated, so reducing personalliability through incorporation seems moreattractive.

As I am sure you understand, business confiden-tiality is very important to me and my associates.We desire to incorporation as confidentially as wecan. Please inform us what documentation andpaperwork is required and how much these serviceswill cost?

I would like to start the process of incorporationas soon as possible. Also, how much can we expectyour fees to be?

Due to numerous professional commitments, Iwould prefer to communicate through email. Ihope to hear from your soon.

Thank you very much, [alias]

Treatment Email: International Standard

Dear [name/company]

I am contacting you regarding a business I am tryingto set up. I am a consultant and my colleagues and Iare seeking to establish an international corpora-tion. I am [Norstralia] resident, but I do businessboth locally and with some international clients,including some in your region. Out business hasbeen growing substantially, and our goal is to limittax obligations and business liability.

We would like as much business confidentialityas possible in these early stages of formation. Myinternet searches show that the internationalFinancial Action Task Force requires disclosure ofidentifying information. But I would rather notprovide any detailed personal information, ifpossible.

So, we would like to know what identifyingdocuments will be required to establish this com-pany. We would also like to know what start-upcosts will be.Due to my travel schedule, email will be the best

way to reach me. I look forward to hearing fromyou soon.

Regards, [alias]

Treatment Email: Consequences

Dear [name/company]

I am seeking information on how to incorporate aninternational company. I hope you might be ableto offer what I need.I am a consultant, and my colleagues and I live in

[Norstralia]. Much of our business originates here,where we operate, but our company also growsquickly among international clients. Many of themare in your area. So, we feel that incorporation is anecessary option of us. We hope to limit taxesobligations and business liability.We would like to know if you feel that you will be

able to service us with a corporation. What iden-tifying documents will you request for this trans-action? We would prefer to limit disclosure as muchas possible.My internet searches show that the international

Financial Action Task Force sets standards fordisclosure of identifying information when form-ing a company. I also understand that legalpenalties may follow violation of these standards.But I would like to avoid providing any detailedpersonal information if possible. If you couldanswer these questions and also let us know aboutyour prices, we very much appreciate it.Thank you for the time to address our query.

Business obligations make communication diffi-cult, so we would prefer to correspond with email.

Until we speak again, [alias]

Treatment Email: Appropriateness

Dear [name/company]

I ama residentof [Norstralia] andwould like to inquireabout your process to form international corpora-tions. With several associates, I operate a consultingfirm in [Norstralia].Wedealwitha growingnumberofinternational clients,many that come fromyour area,and would like to pursue incorporation option forliability and taxes purposes.


618


We are particularly concerned with keeping busi-ness interactions private; thus, we are eager to limitinformation disclosure as much as possible. Myinternet searches show that the international Finan-cial Action Task Force sets standards for disclosureof identifying information when forming a com-pany and most countries have signed on to thesestandards. As reputable businessmen, I am sure weboth want to do the right thing by the internationalrules. But I would like to avoid providing anydetailed personal information if possible.

Can you please inform me what your start-upcosts are and what kind of identification or docu-ments we will need to provide? We are all fairlyburdened with commitments, so email communi-cation is preferable.

Thank you in advance, [alias]

ABOUT THE AUTHORSBrent B Allred is Professor of Strategy and Inter-national Business in The College of William andMary’s Raymond A. Mason School of Business. Heearned his PhD from The Pennsylvania StateUniversity. His current research interests exploreglobal corruption and governance, technologysourcing, and patent rights. His papers have beenpublished in journals such as the Journal of Inter-national Business Studies, Journal of Product Innova-tion Management, Management International Review,Journal of International Management, and Academy ofManagement Executive.

Michael G Findley is Associate Professor ofGovernment in the University of Texas at AustinDepartment of Government. He earned his PhD

from The University of Illinois at Urbana-Cham-paign. His current research interests include politi-cal violence, international development, andinternational law. His work has been published injournals such as Strategic Management Journal, Inter-national Organization, World Development, PublicChoice, Complexity, American Journal of Political Sci-ence, Journal of Politics, University of Minnesota LawReview, and Cambridge University Press.

Daniel Nielson is Professor and Associate Chair ofPolitical Science at Brigham Young University. Hiswork studies international development throughfield experiments, with a focus on corruption andgovernance. He is co-author of Global Shell Games:Experiments in Transnational Relations, Crime, andTerrorism (Cambridge, 2014) and co-editor of Dele-gation and Agency in International Organizations(Cambridge, 2006). Among other journals, hisarticles have appeared in the Strategic ManagementJournal, American Journal of Political Science, Interna-tional Organization, University of Pennsylvania LawReview, International Studies Quarterly, and WorldDevelopment.

J C Sharman is the Sir Patrick Sheehy Chair ofInternational Relations in the Department of Poli-tics and International Studies at the University ofCambridge. Sharman’s research is focused on cor-ruption, money laundering and tax havens, as wellas the international relations of the early modernworld. His latest books are International Order inDiversity: War, Trade and Rule in the Indian Ocean(Cambridge University Press, 2015 with AndrewPhillips) and The Despot’s Guide to Wealth Manage-ment (Cornell University Press, 2017).

Accepted by Douglas Cumming, Guest Editor, on 22 September 2016. This article has been with the authors for two revisions.


619


anonymous shell companies: a global audit study and field ... · anonymous shell companies: a...

Documents