The compliance game: legal endogeneity in anti­bribery settlement negotiations

In line with a wider expansion of global governance tools to prevent and punish corruption, the enforcement of anti-foreign bribery laws is increasing. Yet few cases are tested in court; rather, confidential negotiations and out-of-court settlements have become the norm. Many argue that this favours the enforcement authorities, since even firms with a good defence are reluctant to risk a trial when settlement appears a safer option, but there is very little evidence about what really happens in anti-foreign bribery negotiations and how companies and enforcement authorities interact during this process. Through qualitative research with participants in negotiations for settlements of the United States Foreign Corrupt Practices Act and the United Kingdom Bribery Act, we find evidence that companies have considerable power to influence how the law is interpreted and implemented. In particular, companies use compliance programs to signal good character, while enforcement authorities willingly accept these symbols as justification for settling out of court or acting leniently. Companies therefore play a key role in shaping the law in places where it is ambiguous, an interesting example of ‘legal endogeneity ’ , whereby the subjects of the law help to shape its meaning.

cooperative stance and subject to future good behaviour (Ryder and Pasculli, 2020;Søreide and Makinwa, 2020;Lord, 2018, 2020). Thus, most violators avoid larger fines and the risk of collateral consequences of criminal prosecution of corporations such as debarment from government contracts; similarly, bribery revelations rarely do much damage to companies' ability to operate or win future business (Karpoff et al., 2014;Garret, 2016). Yet it is possible that the impact of these laws occurs through a different mechanism that is more preventive, namely by incentivising companies to reform internal governance and introduce compliance programs, which may have the effect of materially reducing the payment of bribes by employees.
To research this further, it is important to explore the interaction between enforcement authorities and corporations during settlement negotiations. Some scholars argue that the reliance on NTRs creates perverse incentives whereby companies might rather acquiesce to a settlement even if they have a good defence because the risk of being penalised for non-cooperation is too high, making enforcement authorities the dominant party in negotiations (Koehler, 2009;Willborn, 2013). However, there have been few empirical investigations of settlement negotiations, with the literature generally relying instead on official settlement documents, which present a carefully constructed narrative about the final outcome but offer little insight into the nature of the intervening dialogue.
Against this backdrop, we pose the following question: how do corporations influence and shape the meaning of anti-bribery laws? Focusing on settlement negotiations, we collect empirical data through interviews with key participants about how companies and enforcement authorities interact during this process. 1 We find that the gap between the risk implied by the legal framework and the reality of enforcement practices is a major source of ambiguity for companies seeking to mitigate bribery risks. Yet because anti-bribery compliance programs are specified as legitimate grounds for the enforcement authorities to treat corporations leniently, companies respond by investing in compliance programs to signal their 'good character'. This helps firms exercise agency over how law is interpreted and implemented. The enforcement authorities, meanwhile, are averse to the risk of causing corporate failure, and face incentives to rely on compliance programs as evidence of good corporate behaviour which justify leniency. This creates a further reason for companies to invest heavily in these 'symbols' of compliance. Over time, these compliance programs become 'endogenous' to the law: interactions between enforcement authorities and companies during settlement negotiations are dominated by the display and interpretation of compliance symbols, which shape the meaning of an 'effective' compliance program and influence how antibribery law is interpreted and implemented, in line with legal endogeneity theory (Edelman and Talesh, 2011;Edelman et al., 1999;Talesh, 2012). To the extent that compliance programs are effective in changing corporate behaviour, this might mean that anti-bribery laws are reducing the frequency of bribes paid in international business, and therefore that they are having an impact on overall corruption. Our results thus inform our understanding of how public and private policing interact, and provide a nuanced account of the links between enforcement practices, corporate compliance programs, and behavioural change.

Anti-bribery law enforcement and compliance
Historically, governments have taken heterogeneous approaches to corporate bribery (Doig, 1984). While domestic bribery was criminalised in many countries, bribery of foreign government officials by corporations was often unregulated or tolerated because of national economic/political interests (Doig, 2011). Consider, for example, the FCPA, which was scarcely enforced in its first two decades, in part owing to lobbying from US companies who complained that it put them at an unfair disadvantage when competing on international markets (Brewster, 2017;Hock, 2014). But since 1997, when the OECD Anti-Bribery Convention was signed, 38 OECD members as well as six non-OECD countries have committed to introduce similar laws, even if their implementation and enforcement vary considerably. This imbalance of legislating more against domestic bribery than international bribery, thus, may be being steadily reversed .
While in the 2000s and 2010s the enforcement of anti-bribery laws became more active and hence the risk for firms increased considerably, enforcement authorities remain constrained by the potential collateral consequences of their actions. Criminal conviction for violating anti-bribery laws can have very serious implications for firms: in addition to heavy fines, indictment could lead to reputational damage and to companies being debarred from bidding for public contracts all around the world (Acorn, 2021). In the US, prosecutors are very aware that corporate convictions have in the past led to corporate failure (albeit for non-bribery cases such as Arthur Andersen), with major consequences for jobs. As Pollack puts it, "as reluctant as the corporation is to see that result [indictment], the government is only slightly less reluctant" (Pollack, 2009(Pollack, : 1404. In other words, for both the company and the enforcement authorities, there is great pressure to settle rather than test a case in court (Hawkins, 1983).
Indeed, in the vast majority of cases, the DOJ and the SEC use negotiated agreements, rather than going to trial, in order to resolve allegations of foreign bribery with corporate defendants. This trend has also been followed by other countries which are signatories to the OECD Anti-Bribery Convention, and is sometimes driven by resource constraints as well as a host of other reasons (Bismuth et al., 2021;Grasso, 2016;Hock, 2020;Lord, 2018, 2020). The United Kingdom, France, Argentina, Canada, and Singapore have all recently begun to permit prosecutors to defer or drop prosecutions of a firm for corruption offences in return for the accused firm paying a fine, adopting measures to prevent future offences, and cooperating with ongoing investigations.
Corruption is a practice where the perpetrators often have sufficient power to cover their tracks and obscure evidence, creating a problem of asymmetric information: companies typically know more about the wrongdoing than enforcement authorities do and hence the latter rely heavily on self-reporting by the former (Ivory and Hock, 2021). The more difficult it is for prosecutors to accurately observe actual behaviour, the more they rely on information from companies to make judgements. Privately-led internal investigations play an invaluable role in the process of corporate crime policing (Button, 2019), granting power to corporations to frame the information they provide so as to limit the scope of the investigation. This interaction between companies and law enforcement authorities during this process takes place in a context of considerable ambiguity.
The struggle of criminal justice systems to deter bad corporate behaviour has been well documented (Sutherland, 1949;Levi, 2008;Lord and Levi, 2015;Shepherd and Button, 2019). Literature clearly indicates that there are many approaches and mechanisms, which might be perhaps more legitimate and effective than negotiated settlements, used by criminal justice systems to curb corporate bribery (see Lord et al., 2020). The scope of this paper is more limited as we empirically investigate only one specific case study, namely the process of foreign anti-bribery negotiations.

How guidance shapes incentives
Since corporationsand the lawyers that advise themlack judicial precedents against which to estimate how their behaviour might be interpreted, they rely instead on interpreting the many sources of guidance. The US enforcement system grants wide discretion to prosecutors as to whether or not to pursue corporate crime cases. They should decide through reasoned exercise of prosecutorial authority, guided by a series of policy documents, which have evolved in response to waves of regulatory activity (Weber and Wasieleski, 2013: 609-610). The most relevant policy documents for corporate anti-foreign bribery enforcement and settlement negotiations are set out in Table 1.
The US Attorneys' Manual applies to all crimes committed by business organizations. It contains detailed guidance on how to determine whether a compliance program is merely a paper program. The Sentencing Guidelines also apply to all types of corporate crime, but are mainly relevant for the latter stages of an enforcement action. They create an incentive for companies to introduce ethics and compliance programs, to decrease the fine levied or facilitate a deferred prosecution. Chapter 8 of the Guidelines sets out formulae for calculating sanctions, with considerable discounts that reward self-disclosure, cooperation, and remediation. For example, when a company has voluntarily self-disclosed misconduct, fully cooperated, and remediated in a timely and appropriate manner, it can receive a declination, or a 50% reduction from the bottom of the penalty range. Even if a company did not voluntarily self-disclose but later fully cooperated and timely and appropriately remediated, it could receive up to a 25% reduction. Additionally, in 2017, the DOJ issued the Evaluation of Corporate Compliance Programs guidance to assist prosecutors in making informed decisions as to whether, and to what extent, a corporation's ethics compliance program was effective at the time of the offence, and at the time of settlement negotiations. The guidance states that the DOJ will focus on practical operation of corporate compliance programs, mainly as they relate to corporate culture, compliance structure and resources, and the effectiveness of policies and procedures.
The DOJ also issues foreign bribery-specific enforcement policies and guidance. Most importantly, the FCPA Corporate Enforcement Policy, preceded by the FCPA Pilot Program, provides an incentive framework for foreign bribery settlement negotiations. The framework permits either full or limited credit to be awarded for: a) voluntary self-disclosure; b) full cooperation; and c) timely and appropriate remediation.
Ultimately, these sources of guidance do little to reduce the ambiguity. Together they have so many variables that it is difficult for a company to predict what size of penalty or other sanctions might be imposed in the event of a conviction. Whilst coping with this uncertainty, companies have a strong incentive to provide signals to the enforcement authorities that could be interpreted as evidence of robust control systems.

The negotiation process
Where a firm faces allegations of FCPA violations, inductive analysis of our data allowed us to discern three distinct stages to the negotiation process (see Table 2).

Stage 1: entering negotiations
Negotiations can be commenced as a result of a firm making a voluntary disclosure or the government launching an investigation. The FCPA Corporate Enforcement Policy defines self-disclosure as occurring when the company discloses the conduct prior to an imminent threat of disclosure or government investigation. The enforcement authorities also judge the level of effectiveness of a compliance program at the outset of negotiations.
In deciding whether or not to voluntarily self-disclose, companies are in a relatively strong position, because they usually know Unique issues presented in FCPA matters, aimed at providing additional benefits to companies based on their corporate behaviourself-disclosure, full cooperation, and remediation more about their behaviour than the enforcement agency. At the same time, there is considerable risk attached to not self-disclosing, because this automatically means foregoing the associated discounts. 2 This gives prosecutors an important advantage. Stuntz (2003), for example, argues that when negotiating settlements, laws represent a mere menu of options prosecutors may use, including debarment and high sanctions. Without empirical data on how negotiations take place in practice, one might expect that the ever-present threat for the company of being criminally prosecuted weights the balance of power significantly in favour of enforcement authorities.

Stage 2: establishing facts
In stage 2, the facts of the case must be established and corporations have an opportunity to demonstrate what remedial actions they have taken. As discussed, the corporations have an advantage with regard to knowledge of the facts that they may utilize to frame the case in a certain way or try to restrict its scope. However, they also face incentives to demonstrate cooperation including, for example, through timely preservation, collection, and disclosure of all relevant documents and facts gathered during a company's independent investigation. Remediation can include a root cause analysis, implementation of an effective ethics and compliance program, the payment of disgorgement, forfeiture, and/or restitution. In the US, a corporation that does not self-disclose can still show full cooperation and remediation at stage 2, and hence qualify for a discount of up to 25 percent on penalties (and also fight for a preferable type of resolution such as a DPA). Hence stage 2 is an opportunity for a corporation to show that they had a good compliance program or to demonstrate that they are improving their program to prevent future transgressions; these are relevant at stage 3.

Stage 3: agreeing consequences
In stage 3, the negotiations move to determining how behaviour and evidence from stages 1 and 2 should be reflected in the final resolution. This includes agreeing on the legal interpretation of the corporate misconduct, determining the sanctions and any other legal obligations, and deciding the form and text of the resolution. Here, the corporation is interested in minimising its liability and sanctions, and protecting its reputation.

The role of corporations in shaping the lawlegal endogeneity
This section reviews the literature on legal endogeneity and highlights its value in researching the corporate sector, and made the case for its application in relation to the anti-bribery context. Conventional analysis suggests that laws deter bad behaviour when the risks and penalties outweigh the benefits of misconduct (Becker, 1968). Organizations have many motivations for complying with laws, both negative (fear of consequences) and positive (sense of obligation), or classified as economic, social and normative (Nielsen and Parker, 2012). Socio-legal research focuses on a range of motivations for actual compliancethe threat of punishment (Beams et al., 2003), incentives embedded in internal compliance systems (Hess, 2009) or the psychology of behavioural ethics (Nadler, 2020) but also tends to assume that the law is created and enforced autonomously by the state; which is simply not the case.
In some instances, private institutions play a role in formulating or giving shape to law (Landa, 1981). For example, several corporate scandals in the late 1990s led to the adoption of the Sarbanes-Oxley Act of 2002 and catalysed a wave of self-regulatory and legally mandated corporate reforms associated with the expansion of corporate ethics and compliance programs (Levi and Reuter, 2006;Weber and Wasieleski, 2013). Enforcement authorities began to place great importance on such programs, and it became worthwhile for organizations to invest significant resources in creating and implementing them, to reduce their risk and establish an affirmative defence if employees were found to have violated laws .
In this context, legal endogeneity theory is important in researching the corporate sector because it explains how compliance programs become 'endogenous' to the law through interactions between enforcement authorities and companies (Edelman and Talesh, 2011;Edelman et al., 1999;Talesh, 2012). Given the importance of compliance programs in relation to the anti-bribery context, the theory sheds new light on a peculiar case of how corporations influence and shape the meaning of anti-bribery laws.
Having a good compliance policy on paper is not the same as implementing it well; research suggests that organizations systematically fail to embed rules in business practice (Dávid-Barrett et al., 2017;Edelman, 2016;Krawiec, 2003Krawiec, , 2004. This, however, does not mean that ethics and compliance programs have no practical effect. Legal endogeneity theory suggests that, through organizational mediation, the meaning and impact of ambiguous laws is often shaped or 'managerialized' by business organizations (Edelman et al., 1991). For example, in certain areas of corporate law such as gender equality in the workplace, organizations and the professions together construct "rational responses to law, enabled by 'rational myths' or stories about appropriate solutions that are themselves modelled after the public legal order." (Edelman et al., 1999: 406). These managerial conceptions of law eventually emerge in court decisions , legislation (Talesh, 2015(Talesh, , 2012(Talesh, , 2009, and regulation (Bozanic et al., 2012) demonstrating considerable corporate influence (albeit in a subtler way than influence exerted through state capture (Hellman et al., 2003), lobbying (Larrain and Prüfer, 2015), and corporate cooptation (Campbell, 2001). Legal endogeneity theory thus highlights the potential of organizations to influence and shape the meaning of law through symbolic structures and policies, taking our understanding beyond the traditional dichotomy of compliance versus non-compliance. Legal endogeneity is a processual model that illustrates how private actors influence the interpretation and implementation of law on the ground (Edelman and Talesh 2011). This focus on the actual process of interpretation and implementation through which structures become symbolic, rather than the substance of formal laws, is an extension of Edelman's original theory. Gilad (2014: 134) argues: […] organizations respond to legal uncertainty by adopting formal structures to symbolically signal their compliance. These structures, however, tend to embody businesses' managerial and commercial values, as opposed to regulatory goals. Law becomes endogenous insofar as legal actors then defer to businesses' institutionalized ideas about regulation and compliance.
This has been explored in the context of internal dispute resolutions (Edelman at al., Gilad 2010) and arbitration (Talesh, 2012). For example, Talesh, 2012 demonstrates how the structure of dispute resolution shapes the extent to which private values influence the meaning and implementation of consumer protection law, and consequently, the extent to which repeat players are advantaged. Typically, the legal endogeneity process is driven by professionals in corporate human resource and compliance functions who have an interest in exaggerating risk and presenting their own roles as critical to mitigating risk (Dobin and Kelly, 2007;Edelman et al., 2001). These actors therefore co-construct regulation and compliance (Edelman et al., 2011: 890;Gilad, 2014).
We contend that the legal endogeneity approach can enhance our understanding of anti-foreign bribery enforcement, because the reliance on NTRs exacerbates the ambiguity of the law and creates ample scope for its meaning to be shaped by reciprocal interactions between public and private actors 3 as well as their intermediaries (Talesh and Pélisse, 2019). Fig. 1 illustrates the process. The law stipulates that leniency can be applied to companies on the basis of their compliance programs. This prompts corporate defendants to seek to convince the authorities that their compliance programs are effective and to signal that they are a company of good character. The enforcement authorities also find the ambiguity of law challenging and welcome organizations' efforts to structure this space (Shaffer, 2009: 156-157;Dávid-Barrett, 2011). In a context where there is little evidence about what makes an anti-bribery compliance program 'effective', its adequacy is determined during settlement negotiations between enforcement authorities and businesses facing allegations of bribery (Garret, 2016), in turn shaping the law by reducing ambiguity about the conditions in which leniency will be applied.

Data and methods
The core research question is: how do corporations influence and shape the meaning of anti-bribery laws? Our research was designed to encompass both 'law in books' and 'law in action' (Headworth, 2020) and to understand the behaviour of participants in negotiations. The research started with an observation that companies use compliance programs to signal good character, while enforcement authorities willingly accept these symbols as evidence of good corporate behaviour as justification for settling out of court or acting leniently. This led us to the concept of legal endogeneity.
To examine the process of negotiating out-of-court, we first reviewed key features of the FCPA, available policy guidance that aims to mitigate the existing ambiguity of foreign anti-bribery laws, and academic literature. The findings of the desk research informed the construction of an interview schedule that included general questions on the process of settlement negotiations as well as specific questions relating to the interaction between enforcement authorities and companies (see Table 3).
To study how written law and policy are implemented in practice through settlement negotiations, we use a qualitative exploratory design (Guest et al., 2011: 7-8) that allows for flexibility in our effort to discover trends and practices that take place behind the veil of foreign anti-bribery negotiations. (Boeije, 2010: 3). Collecting data about a negotiation process which happens in private, influenced by motivations that are impossible to observe, is challenging. It is essential to gather first-person accounts because documentary evidence is unlikely to contain all of the information needed. However, it is also difficult to gain access to individuals who have been involved in negotiations and to discern whether they are truthfully reporting their experience.
Our strategy has been to conduct interviews with a set of practitioners who have direct and longstanding experience of negotiating settlements and of conducting related tasks such as internal investigations. Data collection ceased when the point of saturation was reached such that new interviews failed to provide significant additional insights. Non-probability sampling was used to ensure a breadth of perspectives from enforcement authorities as well as lawyers who represent corporations during negotiations (Kalof et al., 2008: 44). We collected interview data for the purpose of logical inference, and used sequential interviewing so that each experience provided an increasingly accurate understanding of our research question (Small, 2009). This approach is, by design, not representative, but such careful selection of participants with adequate experience of the phenomenon can generate rich, dense, and focused data (Cleary et al., 2014).
Access was initially gained through trusted contacts in an effort to ensure a willingness to speak frankly, and subsequently benefited from snowball sampling, as respondents recommended other contacts. The first author interviewed 13 individuals (see Table 3) -nine US experts and four UK expertswho together have more than 240 years' experience of negotiating settlements and conducting internal investigations. Such careful selection of participants with adequate experience of the phenomenon can generate rich, dense, and focused data (Cleary et al., 2014;Walsh and Downe, 2006). This process was guided by the principles of appropriateness and analytical redundancy whereby, given previous interview material, one will not provide additional insights (Sobal, 2001).
The semi-structured interviews sought to facilitate in-depth exploration of each expert's experiences about the way that negotiations commence, how the relationship between the Parties develops through the various stages, and how the content of settlements is determined. Our analysis of the FCPA, accompanying guidance, and academic literature on FCPA interpretation and enforcement informed the design of an interview guide, as did the theory of legal endogeneity, which prompted us to pay attention to relevant symbols and how they were deployed. (Kalu, 2017: 46). The interviews were conducted face-to-face in London, New York, Washington DC, and online in spring and summer 2019. All but one interviews were recorded, transcribed verbatim and coded for analytically relevant themes.
This paper is a first step on a hard question of what really happens in anti-foreign bribery negotiations. The paper is a snapshot of anti-foreign bribery enforcement as it primarily focuses on US FCPA. While authors have engaged in discussions with some acting prosecutors, these discussions only informed the direction of the research process and authors were not allowed to include these discussions in the presented research material. Moreover, while the data include UK respondents, their insights are primarily relevant to US negotiations, for example when UK companies were involved in settlements with US authorities. When the UK respondents are speaking to the UK context, this is mainly for comparative purposes and the paper does not aim to provide a true analytical comparison between the UK and the US.

Findings
Our analysis suggests that enforcement authorities and corporations face little incentive to go to trial, and other than in exceptional circumstances, are extremely likely to opt for negotiations aimed at an NTR. At each stage of negotiating the settlement, corporations have multiple opportunities for agency: they strategically disclose and display information about their conduct, policies and practices, and take on responsibilities of self-investigation, in an effort to convince the authorities that their behaviour is cooperative and to treat them leniently.

Stage 1: entering negotiations
Much legal scholarship assumes that the ever-present threat of being criminally prosecuted weights the balance of power in favour of enforcement authorities rather than corporations, since companies have such a strong incentive to avoidor, strictly speaking, 'defer' -a prosecution (Koehler, 2009;Willborn, 2013). Yet many of our interviewees indicated that, in the US, especially, prosecutors are often equally averse to prosecuting corporations, aware that past corporate convictions have led to costly corporate failure. Since Arthur Andersen, the US enforcement authorities have found many ways to prevent collateral consequences.
There are, of course, limits. Criminal-purpose organizations are unable to access this protection for example, as one interview respondent noted, "The prosecutor cannot, you know, offer a negotiation to somebody that is totally criminal, or for facts that are totally different from the reality" (P11). Moreover, once negotiating, parties must act in good faith. Participant 9 recalled one instance where this had not been the case: […] the Department of Justice just must have decided, you do not negotiate in good faith. We've, you know we've been at this a long time. […] in certain circumstances the Department of Justice may well say, if it sees there is not a good faith effort, not a genuine effort to reach a resolution, they can say, that's that and move to indictment. (P9) These rare instances notwithstanding, the US enforcement authorities favour finding a solution that will not cause significant collateral damage and as such, are primed to identify reasons to avoid a prosecution. They can utilize many techniques to prevent collateral consequences, even in cases of endemic bribery, including limiting the scope of the investigation, excluding certain facts from settlement documents, reducing penalties based on the inability of large corporations to pay, and re-interpretation of the legal nature of facts.
When entering negotiations, it is crucial for the firm to establish a reputation for good behaviour and cooperation, and the clearest way to do this is to pre-emptively self-disclose misconduct. Our interviewees generally took the view that self-reporting "gives you a step-in credibility" (P6). Nevertheless, the approach to self-reporting varies. In the UK, corporations must self-disclose wrongdoing at the outset in order to qualify for a DPA while in the US system, the possibility of negotiating a settlement remains open much further into the process. In other words, even if a company does not self-disclose in good time, if they then show willingness to cooperate and change their behaviour, the US authorities are likely to accept this as adequate to move to an NTR.
The key issue in the US is the extent of cooperation and remediation but not necessarily its timing. In other words, even if a corporation only starts adhering to symbolic compliance structures during the second stage of negotiations (see Table 2), it has a very good chance to resolve even the most serious allegations of endemic bribery out-of-court. As two participants explained: […] if they get you to do all the stuff they want you to do and they get a big fine and you say you're terribly sorry they're happy (P1).
[…] it is very much the case that once you get the facts on the table, and you can make a credible argument this is not a criminal enterprise, this is these problems, what do we do to fix these problems, not kill the whole patient? (P6) Each corporation approaches the question of self-reporting a little differently. According to some interviewees, relatively smaller instances of foreign bribery are not worth self-disclosing, and they report them only in part. While this might appear dishonest, or at least lacking in candour, incomplete self-disclosure does not pose a major obstacle to resolving bribery allegations out-of-court, particularly in the US. Nor does 'endemic bribery', where a corporation has paid bribes multiple times in multiple places. Even in cases of endemic bribery involving corporations without serious compliance measures, the US enforcement authorities remain open to negotiation provided a corporation is willing to start cooperating.
Another way of establishing good character is being able to display a robust compliance program. Regardless of whether a corporation self-discloses its conduct, one of the first moves of the enforcement authorities is to assess the corporation's ex-ante ethics and compliance program, i.e., the policies and procedures that existed prior to the alleged violations occurring. Participant 5 described the key mechanics of this process: It starts before the negotiation; it starts with the investigation. And when you're doing the investigation, you should be looking at what kind of compliance program did they have and why did it go wrong. And so, if you find that the only training that happened was the last half hour before the cocktail party […], you know they're not serious about compliance programs. So, you would pick that up during the investigation stage, and then during the negotiations, you would have that as part of your background. (P5) If a corporation is not able to convince enforcement authorities that they had a good ethics and compliance program at the time of the alleged offence, their negotiating position is weakened. A corporation's adherence to symbolic compliance structures helps enforcement authorities to determine what level of trust to place in the corporate defender and its ability to investigate the nature and extent of misconduct.
Historically, many non-US corporations in particular have failed to symbolise compliance adequately at the outset, resulting in very difficult negotiations, high levels of sanctions and greater risk of being criminally prosecuted and potentially debarred. One US defence lawyer explained this phenomenon: […] the major negotiation was in Brazil, and the problem for the Brazilians there was they didn't have a culture of compliance programs, of real compliance programs, so they didn't know how to measure them, and that's I think true … you know, was true in Europe […] We have an informal view of a lot of companies here. (P6) A corporation can also signal good character via the reputation of the private professionals that it hires to represent it. Defence lawyers, compliance advisors, and other specialists hired by corporations often have prior relationships with enforcement authorities and are concerned about their own reputations. To remain trusted and respected, private professionals seek to represent clients that aim to adhere to symbolic structures and remediate during the course of negotiations. As one defence lawyer based in the UK indicated: We wouldn't take a client that we thought was going to commit illegality after we'd taken them on, who wasn't genuinely looking to remedy the problem, but was looking to use us as a cover to commit further crimes, for instance. That's not a situation you want to get in as a lawyer. (P13) Some lawyers are trusted to defend only corporations that aim to negotiate in good faith. Several interviewees confirmed that if enforcement authorities are less familiar with the company and their lawyers, they will likely give the company less space to influence the negotiation process.
Having the right lawyers might even be a way of 'buying influence', in a world in which there is considerable movement through the 'revolving door' between legal, regulatory and corporate positions. Some interviewees argued that the representatives of enforcement authorities and representatives of corporations comprise a pre-existing network of contacts that constantly interact, driven by the wish of some public officials to obtain lucrative positions in the private sector and by the legal representatives' wish to gain access to insider information. It is well documented how such interaction may 'co-opt' public authorities or individual officials (Shaffer, 2009: 156-157). For example, one expert with experience in negotiating settlements on behalf of the UK government explained how this professional network interacts: […] these are very high-level lawyers, they know prosecutors very well, and unfortunately to say in the US, prosecutors would like to become partners to Forbes [list of the top corporations], and there are revolving doors. They know each other, they frequent each other and they go to trial together, so they are, in a way, acquaintances if not friends. So, because of this, these are the people that speak at the beginning, so very very influential lawyers with the prosecutors. (P11)

Stage 2: demonstrating cooperation
In the second stage of negotiations, the Parties establish the facts of the case. This stage is very interactive and iterative, moving back and forth between negotiation and investigation, with corporations seeking to demonstrate willingness to assist the authorities. If the corporation has already credibly signalled willingness to cooperate and remediate in the first stage of negotiations, enforcement authorities will usually invest a high level of trust in the corporations by asking them to undertake investigative tasks on their behalf: […] the company is always going to know more. It doesn't always mean there is lots more bad stuff that they know and are trying to hide. (P10) Similarly, corporations may offer to take responsibility for investigating the scope of the problem as a way of signalling a cooperative stance, leading to a situation where "Most of our US deferred prosecution agreements are not really investigated by the government but by me, by private lawyers" (P6). Indeed, a company also often has more resources than an enforcement authority to investigate the matter. Participant 6 describes one such negotiation with a non-US enforcement authority: […] the facts are in an [region] country. How are you going to get the facts, the documents and the other issues? Oh, said the prosecutor, we have treaties and we have diplomats, it'll only take us two or three years to get the facts, and I said, well, I have a guy with a visa from that country standing by at the airport in [region] ready to go get those documents and have them back here in [region] in a couple of days. He said, oh, you can do that? I said, yes, we can do that.
[…] They were completely in control, but I had a lot more resources than they did, and I had a lot more flexibility because the company's office in that [region] country had the documents and all we had to do was call up the manager from that country and say put them in a box and standby, there's a lawyer coming to pick them up. (P6) Of course, it is always a possibility that the corporation will cheat about what it discloses, but our participants took the view that there was little incentive to do so, because both parties are mainly focused on shaping a resolution, rather than necessarily uncovering every aspect of the wrongdoing.
In cases of endemic bribery, involving multiple bribes paid at multiple times in multiple places, enforcement authorities largely focus on the attitude of the corporation to identifying and fixing systemic compliance issues, rather than seeking detailed identification of every single act. This means that a substantial part of negotiations comprises efforts to limit the scope of investigation. Rather than fight for a maximum penalty, prosecutors are interested in swiftly reducing information asymmetries, defining the scope of the case and securing complete information about what violations have occurred. As explained by a lawyer who negotiated multiple large settlements, both parties need to be flexible in their approach: […] you know, where you're finding problems in every corner there may be less of an imperative to nail down all of them and to a certain extent, where a company has systemic issues, if there's a desire to resolve relatively promptly -and generally our view is a prompter resolution is better -there needs to be some flexibility on the part of the company and the government in terms of where to look. (P8) Once a corporation has started to signal compliance, it has little incentive to hide particular instances of illicit conduct: […] you're either co-operating or not co-operating and if you're co-operating, you've got to do everything you can to cooperate. It's rarely an effective strategy to co-operate … say you're co-operating but then to hold back. You've got to be "all in" in those circumstances.
[…] (P8) Similarly, one defence lawyer indicated that the outset of an internal investigation provides an opportunity to clean the shield of a corporation: […] the goal of the management of the company is to take the cancer out, to have a surgical removal of the cancer.
[…] Asymmetrical information is not the right premise for a DPA and investigation. Asymmetrical information will result in the case not being dealt with. (P6) This explains why, as Participant 4 explains, informational gaps are closed relatively quickly: […] the balance is struck very quickly because all the DOJ and the SEC have to say is, what else? […] there is an incentive to tell them more to avoid them finding out. And if they ask a direct question, we have to answer honestly. So, the asymmetry, in my head I know there's Russia, China and India and when I go in and I start talking about Russia, there is still asymmetry in favor of the company. But if you have a good prosecutor, they ask a series of questions. They will make it necessary for us to tell them about China. They will almost incentivize us to tell them about China and India. (P4) In this stage of negotiations too, a cooperative posture matters more to the authorities than the extent of illicit conduct. The investigation of every instance of bribery is very costly, hence every complex negotiation must grapple with the question of how, and to what extent, the Parties can limit the scope of investigation. This decision will depend on many factors, including type of company, level of trust between an enforcement authority and corporate representatives, and the complexity of a scheme. Ultimately, nevertheless, the Parties will try to agree upon a common definition of what the scope of investigation should be: […] all of this is premised on the notion that there is agreement as to what the real scope is, right. That's always a huge issue. And, how is the scope defined, right […] they are going to engage in a dialogue to really figure out what is the scope, do we really need to go back to 2010 or not. (P10) Participant 4 further describes how flexibly the scope of a case can be determined: They say, okay, thank you for coming in about Russia. We talk about Russia. When we're done talking about Russia we say, the Government will say, where, what other countries? Do you have any similar issues in other countries? And we'll say, mm, nothing as serious as Russia. And then they'll say, what about things that are less serious than Russia? We'll say, yeah, every company has got things going on. And they'll say, like what? So I'm trying to answer their questions honestly, but I'm not trying to give them the information. And if they're a lazy prosecutor, they'll say, okay, so you're just talking about normal things? I'll say, yeah, just the compliance stuff. And then they'll stop asking the question, if they're lazy. If they're hungry, they'll keep going. (P4) In some cases, the scope of an investigation may be narrowed so as to include only the most serious instances of bribery while excluding less serious instances. Enforcement authorities conventionally allow the scope of an investigation to be limited provided the corporation signals compliance and they think the gathered data sufficiently characterize the core of its wrongdoing.
This practice of narrowing the scope of a foreign bribery case further exacerbates the ambiguity of the law for corporations and grants enforcement authorities wide discretion over whether or not to police corporate bribery (Edelman et al., 1991;Willborn, 2013). In large cases, this process results in an agreement about the general nature of a corporation's bribery problem but with only select examples of bribery being thoroughly documented and investigated. For example, if one part of a company that engaged in foreign bribery was active in 100 countries, the corporation may want to thoroughly investigate conduct only in the 20 riskiest countries, and then discuss the findings with the enforcement authorities. As Participant 6 describes: And at the end of 20 countries, the guys in my plan thought they knew what the facts were pretty well, and the government knew, they didn't need all 100 countries […] As such, the scope of foreign bribery casesand the extent to which that scope is narrowed -is up for negotiation: […] everything is negotiable. My client had to acknowledge they'd been paying bribes in [X] of the 20 countries […] and the government understood that this was a good company, the part that we were dealing with, was a good company they didn't want to kill, so they had more than enough to, at the time, charge us with multiple … [X] or whatever those violations […] But they didn't think it was profitable to go to the next [Y] countries. (P6)

Stage 3: agreeing consequences
After the internal investigation has been completed and the relevant facts and circumstances established, the Parties enter a phase of determining the resolution. In this phase, a discussion takes place about the extent to which a corporation exhibited compliance and cooperation. The more and better-quality compliance symbols the corporation has been able to signal, the greater its chance to receive a reduction in penalties as well as other advantages. The Parties do not negotiate only the legal consequences of misconduct, but also how facts should be legally interpreted and which facts and how should be interpreted. They arrive at an agreement about the extent of criminal conduct, and then calculate sanctions in accordance with the mathematical formulae set out in government enforcement policies. Where a company has self-disclosed, or demonstrated cooperation and remediation, specific discount thresholds are applied.
One implication of this approach is that the scale of bribery is typically underestimated. The process is presented as one in which the initial determination of criminal conduct is 'objective', but this underestimates the extent to which that assessment has already been influenced by the corporation, exercising agency through its investigation. The Parties often agreed to limit the scope of investigation before the true extent of criminal conduct was fully known. Cooperative corporations thus in a sense receive a 'double discount'. Unlike in standard economic models, enforcement authorities are not primarily motivated by maximising fines and, for individual corporations, more bribery does not necessarily lead to higher sanctions.
This mismatch between the real scope of a criminal scheme and the 'facts' established during negotiations is further amplified by the corporation's interest in protecting its reputation. Participant 4 explains how the publicly available content of settlement documents does not show the full picture: What you see in the statement of facts is a lot of information, but there's a lot of information that never goes into the statement of facts.
[…] What's included in the statement of facts is usually enough for the Government to show that they have a claim, they have a potential claim. We will say to the Government, you don't need to say more than that. So if your claim is based on facts one, two, three, four, five, we can say, you don't need all five. (P4) Corporations may also suggest an appropriate structure for a case, again playing a role in interpreting relevant laws. Participant 1 indicated that: You're going to give me something for that. But I'll structure this so that [subsidiary in country 1] and [subsidiary in country 2] have to plead guilty. That's terrible for the businesses and the people in those countries, but those are tiny businesses compared to the business in the US for example, the business in Germany, the business in the UKthose guys aren't going to have to plead guilty. Country 1, Country 2, or whatever, they have to plead guilty. (P1) Thus, in deciding how to structure the case, both companies and enforcement authorities are incentivised to focus on 'easy targets' that are not 'too big to fail'.
These practices have implications for the study of bribery in international business. While the Parties negotiate over what facts to publish and how to legally interpret them, and over which instances of bribery to focus on, commentators and academics rely on exactly this information as evidence of the acts of bribery that occurred, and are unable to judge the extent to which the Parties censored the published facts or their legal interpretation. This imposes a serious limitation on the utility of this evidence for academic research. At worst, over-reliance on such documents may lead those analysing the implementation of anti-bribery laws to draw invalid conclusions about the appropriateness of sanctions, the legal qualification of underlying acts, and other important aspects of antiforeign bribery enforcement.

Discussion and conclusion
We have presented evidence about the way that enforcement authorities and companies interact during negotiations to settle foreign bribery cases, and the implications of these exchanges for their effectiveness in deterring bribery. At the first stage of the negotiation process, corporations are eager to signal their cooperative stance and may achieve this through voluntarily self-disclosing, demonstrating a strong ex ante compliance program, or offering to self-investigate. The enforcement authorities wish to avoid a prosecution and hence are interested in recognising these symbols of compliance. The key issues that influence decisions to enter settlement negotiations are not therefore the seriousness of misconduct or even whether a corporation voluntarily self-discloses, but the willingness of corporations to cooperate and remediate. At the second stage, investigations are largely conducted by corporations under the supervision of enforcement authorities and public-private interactions typically lead to the scope of a case being constrained, with only some parts of a bribery scheme being elaborated, potentially excluding many instances of criminal behaviour or weaknesses in systems. The corporation's ability to signal its compliance and cooperative stance in the first two stages also influences the determination of consequences in the final stage. A cooperative firm is effectively rewarded twice, since the extent of its criminal conduct is almost certainly underestimated and it receives penalty discounts for cooperation.
The paper has some important limitations. Besides its narrow focus on anti-bribery settlement negotiations, the data show only limited detail about the complexity of the interests of the compliance lawyers, CEOs, defence lawyers, and other firms. While our findings in section 5.1 discuss the importance of the reputation of the private professionals that firms hire to represent them and the issue of revolving doors, further research should dig deeper into the stances, resources, and strategies of corporate lawyers in the enforcement of anti-bribery laws and reveal how they interact with public agencies and shape the enforcement of laws. Moreover, there remains an important knowledge gap when it comes to the comparison of how these issues function in different sectors and different areas of economic crime (Button et al., 2022).
Despite these limitations, our findings have important implications. First, our study provides unique empirical evidence of how corporate compliance programs have become endogenous to antiforeign bribery laws, in a context where actors are highly incentivised to pursue NTRs. Anti-bribery laws become mere platforms that provide a structure on which the interaction between enforcement authorities and corporations is built. Corporations are invited to signal their adherence to compliance structures throughout the negotiations process to resolve ambiguities created by public orders, thereby co-creating the law through their interactions with public actors. Corporations that adhere to the symbols of compliance are permitted to limit the scope of investigation and claim other discounts and benefits such as lower penalties, favourable legal qualification of their misconduct, and the exclusion of certain facts from settlement documents.
Enforcement authorities tend to infer the absence of culpability from the ability of corporations to swiftly investigate what violations have occurred, explain why their systems allowed them to occur, and address any shortcomings in underlying compliance policies and procedures. Those businesses that obtain the trust of enforcement authorities therefore exercise greater control over the enforcement process. This incentivises corporates to establish complex anti-bribery ethics and compliance programs as an ex ante signal of their stance towards any enforcement that might subsequently arise and helps explain why organizations invest vast sums in such programs. The implication is that the design of anti-bribery and compliance programs may be driven more by the need to create symbols that can be interpreted as evidence of effectiveness than by a wish to reduce the likelihood of bribes being paid.
Second, the nature of these negotiation practices and the endogeneity of compliance programs is shaping the meaning of antibribery laws. Edelman and Talesh (2011: 103) argue that mutual reliance on symbols of compliance results in an institutionalization of beliefs about legality, morality, and rationality which transforms the very meaning of law. In this case, we see that anti-bribery compliance programs designed to mitigate risk become the standard for judging the ethicality and culpability of corporate behaviour. If companies have programs that meet certain conditions, they can avoid responsibility for individual acts of misconduct, even though the occurrence of those acts might well be claimed as evidence that their compliance programs are inadequate. One important factor is, indeed, stealth, which makes detection difficult, thereby increasing the informational advantage of internal staff of corporations in a negotiation. This arguably creates incentives for enforcement authorities to grant more weight to symbolic structures than to contradictory evidence about actual practices, treating policies as a highly credible signal of practice even when it is clear that conduct is not in step with policy .
The impact of this 'reification' of compliance programs on the prevalence of bribery in international business is unclear. Compliance programs may be having a positive effect on corporate behaviour, e.g., by raising awareness of the risks of engaging in bribery, putting in place more robust oversight structures and helping to make the avoidance of bribery a business norm. But equally, the investment in compliance may represent a largely symbolic exercise. Research suggests there is often a considerable gap between compliance policy and practice (Krawiec, 2003), while such programs represent only one part of organizational culture (Weaver and Treviño, 1999). More optimistically, enforcement authorities are requiring more detailed evidence that compliance programs are 'effective', which may prompt companies to engage more meaningfully.
Third, while some scholars consider the lack of court oversight and underlying ambiguity of law as weighting the balance of power in favour of enforcement authorities, we suggest that corporations, particularly large multinational enterprises, generally benefit from this structure. Removing disputes and other legal matters from external control has long been attractive for organizations, as illustrated, for example, by the creation of merchant courts (Landa, 1981). Compliance programs are, in this context, only one element of broader business efforts to exercise agency when laws are ambiguous, ineffective, or even non-existent (Edelman et al., 1999;. The balance of power between public and private actors matters, because the goals of business organizations differ from those of regulators, and hence the balance of power influences whether the law ultimately serves its public policy objectives. Here, further research might build on the work of scholars who examine when, and how, public orders should defer to private ordering (Hadfield, 2017). Leaving the space to private actors may have advantages, for example, as one expert indicated: People in government should not be negotiating those [ethics and compliance programs] unless they have expertise in that area, and in my experience, lawyers often confuse knowledge of the law with knowledge of what works in compliance and ethics programs. (P5) Where the public order specifies what corporations must do in terms of compliance, there is a risk that this results in corporations undertaking the minimum activity necessary to avoid prosecution. On the other hand, a lack of external governance has in many cases resulted in the failure of regulatory regimes, while simply boosting the number of public officials and influential lawyers in the field of corporate compliance (Parker, 2000). The importance of ethics and compliance programs in the emerging global anti-bribery legal regime may be expanding the law's reach by prompting a wide range of corporations to take precautionary measures. Yet, as our research shows, the practices of limiting the scope of cases and applying discounts are having the opposite effect, constraining the reach of the law. The application of legal endogeneity theory to anti-bribery law raises important normative questions as to the extent one wishes to normalise criminal behaviour, and re-ignites the debate as to whether, and when, corporations should be allowed to fail as a result of their crimes.

Declarations
• The authors have no competing interests to declare that are relevant to the content of this article.
• This work was supported by University of Portsmouth.
• This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Faculty of Humanities and Social Sciences Ethics Committee at the University of Portsmouth (February 20, 2019). • Informed consent was obtained from all individual participants included in the study.
• The data that support the findings of this study are available on request from the corresponding author, BH. The data are not publicly available due to their containing information that could compromise the privacy of research participants.