Peer review and compliance with international anti-corruption norms: Insights from the OECD Working Group on Bribery

Abstract How can we make sure that states do not only sign international anti-corruption conventions, but also comply with them once the ink has dried? Peer review among states offers one answer to this question. This article develops a theoretical framework to study the different processes and mechanisms through which peer reviews can contribute to state compliance. It focuses on three processes: transparency, pressure, and learning. The article subsequently applies this framework to the OECD Working Group on Bribery (WGB) in order to identify how far participants in this peer review perceive the WGB as capable of organising these processes, and to what extent they consider these processes relevant for promoting state compliance. Data come from an online survey (74 observations) and 17 in-depth interviews. The findings reveal that this peer review exercise is perceived as effective in creating transparency about state behaviour, mobilising pressure, and stimulating learning. However, the extent to which these processes can promote compliance is more limited. For these processes to work, political will is crucial.


Introduction
How can we make sure that states do not only sign international anti-corruption conventions, but also take action once the ink has dried? Peer review among states offers one answer to this question. Used extensively in the international anti-corruption regime, peer reviews aim to promote reforms in countries by means of mutual intergovernmental evaluations; experts from member states periodically evaluate each other's compliance with international anti-corruption agreements and make recommendations for improvement. 1 While these instruments usually cannot sanction states that disregard their recommendations, peer reviews do hold potential to raise compliance levels in states. Notably, they aim to create transparency about states' performance, mobilise peer and public pressure, and promote learning during the evaluations. 2 Today, many perspective. The article subsequently connects this discussion to the different processes and mechanisms through which peer reviews can affect state behaviour and contribute to compliance.
Empirically, the article offers a systematic, empirical analysis of how transparency, pressure, and learning take shape in peer reviews and can contribute to compliance with international anti-corruption norms. Drawing upon data from a survey and 17 qualitative interviews, it applies the theoretical framework to the Organisation for Economic Cooperation and Development (OECD) Working Group on Bribery (WGB). Since 1999, the WGB has reviewed states' performance under the OECD Convention against Foreign Bribery, the first multilateral legal instrument in the fight against foreign bribery. What makes the WGB an interesting case is that it does not only evaluate states' legal and practical implementation of an anti-corruption agreement. It also reviews compliance, that is, whether states investigate allegations of foreign bribery and prosecute companies that have violated anti-bribery laws. Other peer reviews in the anti-corruption regime, such as the Council of Europe's Group of States against Corruption (GRECO) and the Implementation Review Mechanism of the United Nations Convention against Corruption (UNCAC), do not review state compliance. According to Transparency International, the world's leading anti-corruption Non-Governmental Organisation, the WGB is the 'gold standard of monitoring' in the anti-corruption regime. 10 Hence, this peer review offers a most-likely case to trigger processes of transparency, to exert pressure, and to promote learning about how to fight foreign bribery effectively. If the WGB faces difficulties in this regard, other anti-corruption peer reviews most likely do so as well.
The next section introduces the compliance debate in international law and international relations, and details the theoretical framework. After introducing data and methods (section 2), section 3 outlines the WGB's history and working procedures. Section 4 then investigates the extent to which the WGB is perceived to create transparency of state behaviour, to generate pressure, and to promote learning. The findings demonstrate that the WGB is perceived as effective in doing so. The analysis in section 5, however, reveals that the extent to which these processes can address non-compliance is considered to be more limited. While the three peer review processes can support states in complying with the Convention, political will is crucial for them to be effective.
1. Peer review as a mode of governance More than two decades ago, peer review made its entrance as a new way of governing complex transnational policy problems. Its extensive use in international organisations as diverse as the African Union, the Council of Europe, the European Union, the Organisation of American States, and the United Nations has attracted substantial academic interest. As mentioned in the introduction, several studies have investigated the international conditions under which different peer review processes take shape. Others have sought to explain the domestic impact of these practices in individual countries. However, hardly any research focuses specifically on the relevance of processes that are organised at the international level for state compliance.
This gap in the literature is perhaps unsurprising, as it is difficult to establish a causal relationship between international peer review processes and domestic reforms. Simply because states change their policies in line with a peer review's recommendations does not mean that the review exercise caused this. States might have already intended to implement these reforms for reasons exogenous to the peer review. In addition, states' anti-corruption legislation and efforts are simultaneously under review by various monitoring bodies, which makes it difficult to identify who gave the decisive push for policy reform, if this can be attributed to a single actor at all.
Recognising these limitations, this article does not conduct an objective, external assessment of whether peer review in the WGB has triggered reforms in states. Instead, it analyses the 10 See: {http://www.oecd.org/corruption/anti-bribery/anti briberyconvention/oecdworkinggrouponbriberyininternationalbusinesstransactions.htm} accessed 30 November 2020.
perceptions of OECD secretariat officials and state delegates, a methodological choice that is further explained in section 2. The next section first discusses the concept of compliance and reviews different perspectives on what explains (non-)compliance with international norms. Subsection 1.2 subsequently introduces the theoretical framework, which details the different processes and mechanisms through which peer reviews are expected to contribute to compliance.

Compliance with international law
Why states do or do not comply with international agreements has interested students of international relations and international law for almost two decades. 11 The academic literature distinguishes between treaty implementation, compliance, and effectiveness. Implementation refers to 'the adoption of domestic rules or regulations that are meant to facilitate, but do not in themselves constitute, compliance with international agreements'. 12 It concerns the legal and institutional framework that is designed to put an international agreement into practice. Compliance goes one step beyond implementation. 13 It pertains to the extent to which states also adhere to their international commitments in practice and how far their behaviour falls in line with prescribed norms. 14 In the case of the WGB, this concerns the domestic enforcement of the Anti-Bribery Convention, that is, whether states open investigations into (alleged) bribery cases, start criminal proceedings, and impose sanctions. Finally, effectiveness concerns a peer review's efficacy in addressing a policy problem, such as reducing corruption levels. 15 Absent reliable indicators on the scope of international bribery, a peer review's effectiveness in fighting bribery is difficult to determine. While some countries make data and statistics available on prosecutions and settlements (that is, compliance), we know little about the extent to which this has also deterred multinational companies from paying bribes abroad (that is, effectiveness). 16 Three strands in the compliance literature present contending claims about what determines compliance with international law. The first, the enforcement perspective, takes a rationalist approach and emphasises incentive structures and cost-benefit calculations. 17 States are considered rational actors that will comply with international agreements when the benefits of 11 Simmons, 'Compliance with international agreements ', pp. 77-8. 13 While treaty implementation often precedes compliance, compliance may occur for reasons unrelated to an international treaty. 14 Simmons, 'International law and state behavior'. 15 Oran R. Young, 'Effectiveness of international environmental regimes: Existing knowledge, cutting-edge themes, and research strategies', Proceedings of the National Academy of Sciences, 108:50 (2011), pp. 19853-860. compliance outweigh its costs. The costs of (non-)compliance can be material (for example, the human and financial resources that states need to invest to enforce legislation domestically) or social (for example, damage to a country's international standing). The second strand in the literature, the management school, offers a different explanation for why states comply with international agreements. Managerial theorists hold that non-compliance is often inadvertent rather than a calculated decision. 18 States are generally inclined to comply with the agreements they have signed, but might fail to do so because they lack the capacity or resources to implement an agreement. Inadequate clarity of international commitments, due to ambiguous treaty language, is another reason why states do not always comply. Finally, constructivists add a third perspective to the compliance debate. 19 They contend that states are positively predisposed to compliance with international agreements, as long as these norms are regarded as legitimate. 20 Compliance then becomes the right and appropriate thing to do. 21 As each strand in the compliance literature identifies different reasons for states' (lack of) compliance with international agreements, they also suggest different policy interventions to raise compliance levels. The enforcement literature attaches the most importance to monitoring and sanctioning. By increasing the chances that shirking will be detected, exposed, and punished, these mechanisms can alter states' incentives to defect. Managerial theorists, in turn, assert that sanctions are ineffective at bestand counterproductive at worstin instances where noncompliance is involuntary. 22 Instead, they suggest measures to increase the transparency of states' performance, to promote the interpretation of international rules, and to build capacity. 23 Finally, constructivists emphasise the importance of social learning and socialisation processes. Through social interactions in international environments, involved actors might change their beliefs about appropriate behaviour, readjust their preferences, and internalise new norms. 24 Table 1 summarises the three perspectives on compliance with international norms.
While these three strands in the compliance literature start from different theoretical assumptions, several scholars argue that they should not be seen as competing or mutually exclusive, but that they can complement each other. 25 In fact, combining different compliance strategies and instruments might yield more effective compliance systems. 26 As will be discussed in the next section, peer reviews hold considerable potential in this regard.  Simmons, 'Compliance with international agreements'. This is additionally strengthened by the normative belief that foreign bribery is unethical, and that states should therefore comply with international legislation that criminalises this offence.  Tallberg, for instance, puts forward the proposition that 'compliance systems become most effective when combining enforcement and management strategies' ('Paths to compliance', p. 629). See also Borzel et al., 'Obstinate and insufficient', p. 1383 and Versluis and Tarr, 'Improving compliance', p. 332.

Theorising peer review and its effects
This section proposes a theoretical framework that details the different processes and mechanisms through which peer reviews are expected to contribute to compliance. The article focuses on three processes in particular: (1) creating transparency; (2) mobilising pressure; and (3) organising learning. Before discussing each of these processes and their corresponding mechanisms, it is important to note that not all of them link neatly to a single theoretical school. Instead, they speak to different theoretical assumptions, and matter for compliance in ways that also go beyond the mechanisms discussed in the compliance literature. Furthermore, the three peer review processes do not work independently of each other but can be mutually reinforcing. Transparency, for example, might not only be relevant in itself, but also be a precondition for processes of peer pressure, public pressure, and learning to take shape.
The first process through which peer reviews can contribute to compliance is by creating transparency of state behaviour. 27 Monitoring fulfils an important information function. 28 By periodically reviewing states, peer reviews provide compliance information, facilitate reciprocity, and reassure states that their peers are respecting their commitments as well. This seems particularly pertinent in collective action problems such as the fight against foreign bribery, which gives companies in non-enforcing states a competitive edge over their counterparts in economies that strictly enforce anti-bribery legislation. To break this prisoner-dilemma type situation, 29 the WGB seeks to create openness of state behaviour towards other states and, as such, raise the costs of shirking. Peer review might also increase the costs of non-compliance by informing domestic audiences about their governments' policy performance, 30 which, as discussed below, relates to public pressure.
Peer reviews usually cannot sanction states that make insufficient progress or that ignore their recommendations. Instead, they aim to raise compliance activity by praising active enforcers as well as by naming and shaming underperforming states, which can create peer pressure and public As to public procurement, all companies and states would benefit if no company bribes the contracting party. This means that all states would have to comply with the OECD Convention. However, insufficient insight into the efforts other states make to detect, expose and prosecute foreign bribery encourages defection. pressure. 31 These constitute the second process through which peer reviews are expected to contribute to compliance. Peer pressure might do so by signalling that a country's international standing is at risk. This can create social as well as material costs, as states' inaction might antagonise their strategic partners and harm business relations. Peer reviews can also affect state behaviour in more indirect ways, for example, by informing and mobilising state-and non-state actors that do not participate in the peer review. Specifically, peer reviews provide compliance information to domestic audiences (for example, civil society organisations and the media), who can then hold governments accountable for their actions and exert pressure to comply.
A third way in which peer reviews can promote compliance is by stimulating learning and capacity building among the individuals that participate in the evaluations. 32 More precisely, peer reviews can clarify international obligations through dialogue and give states insight into their policy shortcomings and blockages. They can additionally provide administrative support, foster international networks, and promote the exchange of best practices among national experts. 33 Hence, they hold potential to produce effects in states where non-compliance is inadvertent or due to capacity constraints, which speaks to a management approach to compliance. In line with a constructivist perspective, peer reviews might also encourage social learning during plenary sessions and review exercises. Social learning describes a process in which individuals (in this case: participants in a peer review) change their preferences and normative beliefs through social interactions with colleagues from other states. Consequently, state delegates might be persuaded (rather than forced) to comply with the Anti-Bribery Convention. Table 2 connects the different peer review processes, mechanisms and theoretical perspectives, which will be operationalised in section 2.

Data and methods
To study the different peer review processes, this article combines data from an online survey, conducted in 2015, and 17 qualitative interviews, conducted between 2014 and 2020. Using  While public pressure is expected to change state behaviour by raising the costs of non-compliance, this process takes place outside of the peer review. Thus, a peer review can facilitate or mobilise public pressure (for example, by providing compliance information to domestic advocacy groups) but it cannot exert this pressure itself. the software program Qualtrics, I sent the survey by email to all but two of the 184 state delegates and ten OECD Secretariat members who were involved in the WGB between January 2014 and June 2015 (N = 192). 35 This means that I did a full census (coverage rate of 98.9 per cent). About half of these delegates work for the Ministry of Justice, the Public Prosecution Service, or the Department of the Attorney General. Others work for the Police (including police prosecutors), the Ministry of Economics, Trade and Business, the Ministry of Finance, the Ministry of Foreign Affairs, or dedicated anti-corruption agencies. In these roles, delegates often share responsibility for implementing and enforcing anti-bribery legislation and, consequently, can provide first-hand accounts of the (structural) impediments they experience when doing so. While it is politicians and political leaders that sign up to the Convention, 'ultimately it is law enforcement agencies that implement [it] and take action'. 36 I therefore did not send the survey to ambassadors and permanent representatives to the OECD in Paris, who are often not substantively involved in the country reviews and are not responsible for implementing and enforcing the Convention. I also did not include actors who are external to the WGB, such as businesses, civil society organisations, and the media. Although they play an extremely important role in the fight against foreign bribery, non-state actors are absent at WGB meetings. 37 In total, I collected 74 responses (response rate of 38.5 per cent). 38 The completed sample of respondents reflects the total population of WGB participants in terms of regional background. 39 When it comes to professional affiliation, eight respondents are OECD staff members and 66 respondents are state delegates. Considering the sample size, the empirical analysis consists of descriptive statistics.
As discussed in section 1, the three processes (transparency, pressure, and learning) do not work independently of each other. Transparency might promote compliance by itself but also facilitate peer pressure, public pressure, and learning. To simplify matters, the empirical analysis operationalised the different peer review processes as shown in Table 3. All processes were studied by means of the survey and interviews. However, social learning proved difficult to study empirically; to conclude that social learning has taken place, individuals' norms need to have discernibly changed during a given timeframe. As a longitudinal analysis is outside of this article's scope, the survey probed into respondents' legitimacy perceptions of the instrument of peer review. Interviews additionally reveal how far personal relations and networks have developed within the WGB, and to what extent delegates share the same values.
The formulation of the different survey items was based on exploratory interviews with participants in anti-corruption peer reviews to ensure that these formulations resonate with language used in peer reviews. The survey was self-administered, fully anonymous, and confidential, which may reduce social desirability bias and promote self-disclosure. 40 It also did not force respondents to select an answer. When respondents indicated not to know the answer to a specific survey item, I treated this as item non-response.
Pearson product-moment correlations and Cronbach's alpha were used to study the relationship between these survey items. 41 While the WGB's perceived capacity to generate peer pressure 35 Two email addresses bounced. 36 Interview 14. 37 The OECD does hold annual consultations with business, trade unions, and civil society, and meets with them during country visits. 38 Two reminder emails were sent. 39 Thirty-nine responses from the Western Europe and Others Group of States, 15 responses from African, American, and Asian states, and 13 responses from Central and Eastern Europe. Seven respondents did not disclose their regional background. Western Europe and Others refers to the UN classification and includes Western Europe, Australia, Canada, Israel, New Zealand, Turkey and the United States. correlates with its capacity to exert public pressure (r = .0.432; p < 0.001), their internal consistency is too low to merge the two items (Cronbach's alpha of 0.599). In addition, most dimensions of learning and capacity building (accurate reports, 42 mutual learning, technical assistance, and international cooperation) significantly correlate (Cronbach's alpha of 0.693).
Next to the survey, 17 qualitative interviews were conducted. These interviews help to contextualise the survey findings and to explore to what extent respondents consider these processes relevant for promoting compliance. Two interviews were conducted with staff members of the OECD Anti-Corruption Division and 15 with state delegates (see Appendix). The interviews, which lasted between thirty and ninety minutes, were done in person or by phone, Skype, and Zoom. Interviewed delegates come from different world regions in proportions that largely reflect the total population of participants in the WGB. The study includes interviewees from countries with different levels of enforcement of the Convention as well as long-term members and relative newcomers to the WGB.
This article studies how far peer review is perceived to create transparency, to exert pressure, and to stimulate learning, and how far these processes are regarded as relevant for promoting compliance. This focus on perceptions offers several advantages compared to an external assessment of these processes. As discussed in section 1, it is extremely difficult to establish a causal Theoretically, perceptions of the accuracy of reports fit with both transparency and policy learning. As I found significant correlations between perceptions of accurate reports and various subdimensions of policy learning, but not between perceptions of accurate reports and transparency, I treat this as a subdimension of learning. connection between a peer review and state compliance. An analysis of perceptions helps to unpack the (perceived) relevance of different processes in promoting compliance, by identifying how these processes take shape, how state delegates experience them, and how far they help (or do not help) delegates in overcoming barriers to compliance. That said, this article can make no claims about causal relations. Where possible, however, accounts from interviewees were verified against information from country evaluation reports and public statements.

Peer review in the WGB
The WGB is responsible for reviewing state compliance with the 1997 OECD Convention on Combating Bribery of Foreign Public Officials in International Business Transactions and the 2009 Anti-Bribery Recommendation to further combat foreign bribery. The origins of this agreement can be traced to the neoliberal agenda promoted in the early 1990s, which identified the detrimental consequences of foreign bribery for free trade. 43 While United States (US) companies had been facing restrictions on foreign bribery since the enactment of the Foreign Corrupt Practices Act in 1977, many European countries not only condoned foreign bribery, but even encouraged it through tax benefits. After years of US lobbying, the Anti-Bribery Convention became the first multilateral agreement to level the playing field in international business transactions by criminalising the supply side of bribery.
The WGB has a dual focus on generating pressure and facilitating mutual learning. As stated on its website, the group aims 'to improve the capacity of Parties to fight bribery in international business transactions by examining their undertakings in this field through a dynamic process of mutual evaluation and peer pressure'. 44 It additionally serves an 'educational function as participants discuss problems and different approaches'. 45 Over the past decades, the WGB has grown from 34 to 44 states, which include all 37 OECD members, Argentina, Brazil, Bulgaria, Costa Rica, Peru, Russia, and South Africa. The group meets four times a year at the OECD headquarters in Paris to discuss country evaluations and to exchange experiences on how to enforce the Anti-Bribery Convention. Meetings are held in a closed setting in the expectation that delegates will express their views openly. 46 There are four phases of peer review within the WGB. These each address a particular stage of compliance with the Convention and progressively place more stringent demands on member states. The main objective of Phase 1 is to evaluate the adequacy of national legislation to implement the Convention. This phase comprises both self-and peer assessments in the form of questionnaires, but does not include an onsite visit by the reviewing team. Phase 2 deals with the actual implementation of the Convention. It includes an evaluation questionnaire and an onsite visit of about a week, undertaken by experts from two countries who act as lead examiners, supported by the OECD secretariat. Phase 3 is the first in a series of follow-up phases to the evaluations carried out under Phase 2 and focuses on countries' enforcement of the Convention. It reviews whether states conduct investigations into bribery cases and penalise financial malfeasance abroad. Phase 3 includes onsite visits and questionnaires, but is more focused and shorter than Phase 2. Phase 4 was launched in 2016. It continues its focus on domestic enforcement but additionally reviews several cross-cutting and country-specific issues. 47 Each evaluation ends with the adoption of a country report with recommendations for improvement, which is published on the OECD website. One strength of the WGB is that the state under review cannot block the adoption of the report or its recommendations. In addition, the WGB takes a graduated approach to reviewing state compliance; states cannot proceed to the next review phase unless they have adequately implemented the recommendations from previous rounds. Should their performance be deemed unsatisfactory, the WGB might demand that they do the evaluation for a second time (that is, a -bis evaluation) or even a third time (-ter evaluation).
When it comes to compliance with the Anti-Bribery Convention, results are mixed. Although 44 countries, which together account for more than 66 per cent of world exports, have signed and implemented the Convention, only a handful enforce the Convention domestically. An OECD publication concludes that 21 out of 44 States Parties have never concluded a foreign bribery case. 48 Many states still show inadequacies in their legal frameworks. Progress reports by Transparency International additionally reveal that enforcement of the Convention is erratic; between 2018 and 2020, the number of states that were actively enforcing the Convention declined from seven to four. 49 At the same time, six countries have beefed up their efforts since 2018. While several interviewed delegates admitted that more could be done to prosecute and sanction foreign bribery offences, many of them pointed to discernible results both in their own countries and other WGB states. 50 Examples include legislative changes (for example, in relation to criminal liability of legal persons), legal proceedings against multinationals, and public reopening of investigations.

Transparency
To what extent is the WGB effective in organising transparency, pressure, and learning? When it comes to creating transparency, the survey indicates that a large majority of respondents agree (50.7 per cent) or even strongly agree (33.3 per cent) that in the absence of the WGB it would be difficult to acquire information about the steps that other countries have taken to fight foreign bribery ( Figure 1). 51 Considering that most states do not make a lot of data on enforcement publicly available, the WGB functions as an important platform to exchange information. In the evaluations, the WGB specifically requests states to provide information on (ongoing) investigations and prosecutions of foreign bribery, which are subsequently circulated within the Group. 52 Important information is additionally shared during the roundtable discussion, the so-called Tour de Table,  Since 2010, the OECD has published enforcement data on an annual basis. The questionnaires used in Phase 3 and 4 request states to provide data on domestic enforcement. 53 Interviews 8, 13, 15; another delegate mentioned that you can get a very good insight into states' enforcement actions by attending plenary sessions (interview 16).

Survey question:
-To what extent do you agree or disagree with this statement: 'In the absence of the OECD WGB, it would be difficult to acquire information about the policy practices of other member states in this field'? (N = 69).

Pressure
The second way in which the WGB can promote compliance is by mobilising pressure on states, which can take the form of state-to-state pressure or pressure from the public. Figure 2 shows that more than three-quarters of respondents (76.1 per cent) perceive that the WGB exerts peer pressure to a large extent or completely. 54 At 51.4 per cent, this percentage is significantly lower when it comes to the WGB's capacity to mobilise public pressure. 55 Delegates exert peer pressure by exposing, questioning, and shaming recusant states, holding each other accountable for their performance. 56 Crucially, monitoring is not a one-off exercise that allows states to regress as soon as the review is over. Rather, states have to regularly report on legislative changes and investigations, even when they are not currently under review. 57 Delegates then have ample opportunity to ask each other critical questions; for instance, about why they have not acted on specific foreign bribery allegations, or have not responded to requests for mutual legal assistance. In case of persistent gaps in implementation and compliance, the WGB may request states to do a repeat evaluation or suspend them from advancing to the next review phase. Because of the peer review's ability to take more stringent measures over time, pressure in the WGB 'builds up over different plenaries. Sometimes, … over years, even.' 58 Various interview accounts suggest that peer pressure also works indirectly through states' self-perceived identity and reputation, and does not always need to take the form of naming and shaming. 59 As one delegate remarked: 'from a country perspective, we don't want to be seen inferior or less effective than [member state X]'. 60  If pressure behind closed doors does not suffice, the WGB may threaten to take additional exhortative measures. 61 To mobilise public pressure, for instance, the WGB can contact the minister of the respective country or dispatch a technical or high-level mission. 62 In recent years, such missions have been sent to Argentina, Brazil, Japan, Russia, and Sweden. Further, to shape public opinion and create domestic awareness, the WGB also does not shy away from publishing critical press statements, which are often translated in the reviewed state's local language. 63 Such statements make the evaluation reports available to the public at large, and are regularly picked up by civil society and the media. 64 Concerned about their reputation and image, delegates try to fend off criticism 65 and influence the final wording of these communications. 66 As one of them remarked: 'Sometimes, we spend more time on negotiating the language of the press reports than on the [evaluation] reports themselves.' 67 Even though peer pressure is formally targeted at states and not at individual delegates, the latter often feel the pressure personally. 68 A large majority of surveyed delegates agree (72.1 per cent) or even strongly agree (18 per cent) that they feel under pressure from the WGB to act upon the recommendations they have received (Figure 2). 69 In the interviews, several delegates indicated to feel nervous about their country review or having their reports discussed in plenary, 70 and spoke about the social discomfort they felt when being subjected to review. The atmosphere at these meetings was then also described as 'tense', 71 at times with a lot of 'emotion in the room'. 72 Figure 2. Perceptions of pressure, in percentages.

61
Referring to the discussions on their country's performance, one delegate explained: 'There is challenge and there is critique and there is quite a lot of aggression. My personal view is that it got slightly out of hand … but it was a key part of the dynamics. Sociologically that worked.' 73 Survey questions: -Generally speaking, to what extent do you believe that the OECD WGB: • exerts state-to-state (peer) pressure? (N = 71) • exerts public pressure? (N = 70) -To what extent do you agree or disagree with the following statement: 'I feel under pressure from the OECD WGB to implement review recommendations'? (N = 61)

Learning and capacity building
What sets peer reviews apart from traditional monitoring bodies is their focus on policy learning, capacity building, and international networking. This taps into four aspects: the extent to which the WGB: (1) provides an accurate overview of reviewed states' performance; (2) promotes mutual learning; (3) provides (technical) assistance; and (4) facilitates international cooperation. In total, 48.6 per cent of survey respondents claim that the WGB delivers an accurate overview of the reviewed state's performance to a large extent, and 23.6 per cent hold that the WGB does so completely (Figure 3). 74 Many interviewed delegates are of the opinion that the WGB provides a comprehensive, detailed assessment of states' performance. 75 The fact that evaluation reports draw upon a variety of sources, collected from both state and non-state actors, positively shapes delegates' perceptions of the accuracy of reports. 76 In addition, peer review reports go through various rounds of discussion, which allows all of the involved actors to share their views. 77 Finally, the OECD Secretariat was often lauded for providing additional information in the reviews, thereby contributing to the quality and consistency of evaluation reports. 78 There is, however, no reason for complacency yet; 27.8 per cent of respondents indicate that the WGB provides an accurate assessment of states' performance only to a limited extent.
The findings additionally indicate that the WGB is rather effective in promoting mutual learning and international cooperation. Just under 70 per cent of respondents report that the WGB is to a large extent or even completely able to facilitate these processes. 79 Perceptions of the WGB's capacity to provide (technical) assistance to states are, however, significantly lower. 80 Interviewees discussed different stages of the review process at which these processes occur: Participation in plenary discussions and in the evaluation exercises offer important opportunities for learning, exchange of best practices, and networking. 81  Interview 2, 10, 11. Interestingly, peer reviews that allow for comparably less interactive discussion, such as the World Trade Organisation's Trade Policy Review Mechanism, produce less critical evaluations and more tangential statements by member states; see Karlas and Parízek, 'The process performance of the WTO Trade Policy Review Mechanism'. 78 Interviews 2, 8, 11, 17; interviews 9 and 14 also discuss the importance of the OECD Secretariat for the quality, consistency and effectiveness of the review exercises. law enforcement officials additionally exchange experiences on how to enforce anti-bribery legislation domestically. This helps to build capacity and to form networks to investigate foreign bribery cases. 84 Finally, the mere practice of bringing together different departments to fill out a self-assessment questionnaire offers a key learning moment, and helps to raise awareness domestically about the need to prosecute foreign bribery. 85 This is important, as effective implementation and enforcement requires action from a variety of state actors (for example, different law enforcement agencies and ministries), some of which are less cognisant of the WGB and its work.
Next to policy learning, interviews provide some evidence of social learning. Already in the early 2000s, Kenneth Abbott and Duncan Snidal described the prosecutors in the WGB as a 'nascent epistemic community', 86 an international network of professionals that share similar values regarding foreign bribery. Almost twenty years later, this description still holds. Meeting every three months for a week, delegates develop close networks and personal relationships, building mutual trust and understanding. 87 Through these interactions, values and expectations converge about appropriate behaviour not only in the context of peer reviewing, 88 but also in relation to providing mutual legal assistance and information in transnational bribery investigations. 89 Delegates that do not conform to these expectations risk being shunned or 'mildly ostracised'. 90 In one reported instance, recalcitrant delegates from an (at the time) new member state were simply told that 'this is not how we do things here'. 91 The survey additionally indicates that peer review is widely accepted as a monitoring instrument in the anti-corruption regime: 45.9 per cent of respondents consider its use appropriate and 51.4 per cent even very appropriate. 92 While this does not confirm that social learning has taken placeparticipants might have already held these views before they joined the WGBit does show that peer review is now widely perceived as a legitimate policy tool.

Survey questions:
-Generally speaking, to what extent do you believe that the OECD WGB: • presents an accurate overview of reviewed states' performance ? (N = 72) • triggers mutual learning? (N = 72) • provides (technical) assistance to states to implement the OECD Anti-Bribery Convention? (N = 70) • facilitates international cooperation? (N = 70) -How appropriate or inappropriate do you find it that peer review is used to assess states' performance in the field of corruption? (N = 74)

Peer review and compliance with the Anti-Bribery Convention
The previous section concluded that peer review in the WGB is perceived as quite successful in creating transparency, organising pressure, and stimulating learning and international cooperation. And yet, as shown in section 3, compliance levels are mixed and erratic. Drawing upon interviews, this section explores to what extent these processes are perceived to contribute to compliance.

Transparency
To what extent is transparency considered relevant for promoting compliance? The preceding analysis showed that the WGB fulfils a pivotal role in providing states with information about their peers' performance. Even though in many states the number of prosecutions is still too low to be considered (fully) compliant, some countries are making great strides to comply with the Convention. Delegates from other states notice and recognise these efforts. 93 As already anticipated in section 1, it is extremely difficult to isolate the (perceived) impact of transparency on compliance levels from other peer review processes, such as peer pressure and public pressure. For peer pressure to take shape (both through institutionalised channels and more informally), insight into other states' performance appears to be crucial. 94 Likewise, public pressure can only materialise when a peer review creates transparency about states' performance vis-à-vis domestic audiences. Thus, transparency can be best conceived of as a precondition for these other processes to take shape, which, in turn, might boost compliance levels.

Pressure
To what extent are processes of peer pressure and public pressure regarded as relevant for promoting compliance? Starting with peer pressure: peer review in the WGB was expected to contribute to compliance by signalling that states' international standing and reputation are at risk, creating social and material costs. States that do not fulfil their obligations under the Convention or are uncooperative in transnational bribery investigations might risk alienating their strategic partners. As one delegate explained: 'They are your economic partners who sit there, so it is quite unpleasant if you end up in an argument …. It is not just reputation, but just … you will be perceived as some sort of banana republic.' 95 Recalcitrant behaviour is perceived to make it less likely that other states will respond positively to requests for mutual legal assistance, 96 and might affect diplomatic relations with other states. 97 This corresponds to research by Rochelle Terman and Erik Voeten, who find that strategic partners are less likely to criticise each other in a peer review but, when they do so, are more receptive to this criticism. 98 Peer pressure was, however, most frequently discussed in relation to the threat of having to repeat an evaluation round (that is, the -bis or -ter evaluations) 99 or that other measures, more visible to domestic audiences, will be taken. Repeat evaluations are time-consuming, burdensome, and distract a large number of officials across departments and ministries from their day-to-day activities. 100 As one delegate remarked: A -bis evaluation would be 'the very worst of the bad things that could happen to us'. 101 In several reported instances, just the threat that such measures might be taken has motivated states to make at least minimal (albeit not always sufficient) efforts to comply. 102 Yet, the ability of a peer review to produce reform through peer pressure should not be overstated. As shown in section 4, while delegates often personally feel under pressure to implement reforms, their colleagues who work for other departments or ministries (and who do not participate in the WGB) at times feel less bound by the compromises that have been reached in the OECD. 103 One delegate explained how some ministries are more cooperative in the evaluations than others: The Ministry of Justice might be helping, 'because they have to sit at those meetings. … But then [the Ministry of] Finance is much more distanced from this. Or the … Association of Accountants, they may have heard from this once, but try to make them feel the pressure. That is difficult.' 104 Peer review might raise the material and social costs of non-compliance for delegates who do not want to return to the next WGB meeting empty-handed. This peer pressure, however, does not necessarily transpire to other departments, ministries, or the government. 105 Considering that peer pressure does not always trickle down to all responsible actors, public pressure holds more potential to generate reform, as the material and social costs of noncompliance fall upon governments rather than individual delegates. Interviews indicate that many (although certainly not all) states do care about their reputation, 106 and that delegates try their best to avoid negative publicity. 107 Several delegates discussed how non-compliance might damage the reputation of a country's business climate and foreign investments. 108 In the past, the WGB has threatened to blacklist UK companies for the country's lack of compliance with the Convention, 109 which resulted in a complete overhaul in legislation and the adoption of the UK Bribery Act. 110 Furthermore, in several instances, the high-level missions that have been 97 Interview 8; This also works the other way around: delegates from countries with strong economic, historic or political ties, at times try to assuage the recommendations their strategic partners receive; interviews 2, 8, 9, 10, 12. 98 Terman and Voeten, 'The relational politics of shame'. 99 Interviews 1,3,4,8. 100 Interviews 3, 8; the financial and administrative costs of having to repeat an evaluation round are also discussed in interviews 1, 4. 101 Interview 9. 102 Interviews 1, 3, 9; interview 3, for example, explained how the Public Prosecution Service stepped up their game to be able to present positive results in the WGB, not only to avoid reputational damage but also to make sure they did not have to repeat an evaluation round. 103 Interviews 1, 2, 7, 8, 9. 104 Interview 2, also 1. Another delegate discussed similar dynamics, but mentioned that this might be due to lower awareness of the WGB rather than active resistance (interview 17). 105 The WGB is currently discussing options to take additional measures, which would further raise the material costs of non-compliance (interview 13). 106 Interviews 3, 4, 15. 107 Interviews 2, 3, 13, 14.
108 Interviews 1,3,8,12,13,17; compliance signals that a state's companies can be trusted and that it is safe to invest in them; interview 14. dispatched to states, and in which WGB representatives meet with ministers and other senior officials, are claimed to have set things in motion and to have raised awareness of foreign bribery. 111 While the threat of public shaming may have motivated some change domestically, the impact of public pressure is limited by the perception still present in some WGB states that foreign bribery does not directly affect the public, 112 or that corruption is not a large problem to begin with. 113 In many states, the focus of the public, the media, and politicians on this crime is low. As one delegate explained: 'we informed the government that we had received a negative evaluation report. … It hardly caused a stir, which makes things even more difficult for me, trying to [get people to] implement these recommendations.' 114 External factors, however, such as the exposure of a corruption scandal can have a profound impact on the WGB's capacity to mobilise peer and public pressure, and promote compliance. 115 Well-known examples are the abovementioned UK Bribery Act, which was adopted after a bribery scandal, and Germany's increased enforcement activity after a corruption scandal that involved electronics giant Siemens. Suddenly falling under increased pressure and scrutiny from domestic advocacy groups, the media, and the WGB, states become incentivised to take action against foreign bribery. In addition, governments become more susceptible to adverse publicity on their anti-corruption strategies and more receptive to new policy ideas. As one delegate contended: all we have to do '[is to] hope for a large foreign bribery scandal. Then you will see how fast [governments] will [implement change].' 116 One strength of peer reviews is that, when a scandal is exposed, they have already identified the legislative shortcomings that need to be addressed, offering tailor-made solutions that can be readily implemented. Delegates discussed how the WGB helps to legitimise reform domestically and empowers reform-oriented actors vis-à-vis less supportive government departments and the private sector. 117 They often refer to the OECD reports and recommendations to promote and justify policy reform at home. 118 In line with observations by Tony Porter and Michael Webb on other OECD peer reviews, 'national officials can use the support of their peers in other states to try to sway public opinion in their bureaucratic battles'. 119

Learning and capacity building
Finally, to what extent are learning, capacity building, and international networks considered relevant for promoting compliance? Section 4 showed that delegates learn from their participation in the WGB, and obtain new insights and skills on how to investigate and sanction foreign bribery. This might be particularly pertinent for experts from new member states who may have never investigated or prosecuted foreign bribery before. 120 The WGB additionally creates awareness among law enforcement officials to prosecute this crime, and fosters networks among them that are pivotal for conducting transnational investigations. Within these networks, norms and expectations converge about appropriate behaviour, providing some evidence of social learning. 111 Interviews 15, 17.
112 Interviews 2,8,9,10,11,12. 113 Interviews 4 Yet, these newly acquired skills and networks are partially inconsequential, when insufficient human and financial resources are made available to put them into practice. 121 While none of the interviewed delegates reported to find their obligations under the Convention unclear, capacity and resource constraints pose significant challenges. 122 Cross-border investigations can take a long time and are often complex, 123 while in some states personnel working on foreign bribery are limited, 124 law enforcement agencies are underfunded, 125 and the workload for prosecutors and judges is high. Due to limited resources, choices have to be made as to which alleged foreign bribery cases will be investigated. 126 This poses particular problems in times of economic downturn, when the amount of public funds that can be put towards detecting, investigating, and prosecuting foreign bribery decreases. 127 It would be premature, however, to conclude that non-compliance is primarily due to resource and capacity constraints. Interviews suggest that in many countries foreign bribery is not a political priority. 128 Several delegates mentioned how foreign bribery has moved down on their government's priority lists, 129 making it more difficult to mobilise support and resources to investigate suspected foreign bribery cases. As one of them summed up the political attitude to foreign bribery: 'there is no interest in it, it is not a priority, and there are other things that have to be done'. 130 Rather than spending resources, time, and energy on fighting a crime that is often not visible to the public, several governments focus on tackling other forms of crime, instead. 131 In many Phase 3 evaluation reports, the WGB then also demands that states make adequate resources available to proactively fight this crime. While peer review processes aimed at learning and capacity building can contribute to compliance, there has to be some degree of political will for them to be effective.
profile to foreign bribery, many countries with sizable and thriving export markets do not actively enforce the Convention either. Research by Jo-Anne Gilbert and J. C. Sharman shows that in the past, several governments have resisted compliance with the Anti-Bribery Convention, driven by electoral incentives to (appear to) protect business interests. 135 As long as several large, exporting economies do not prosecute companies that pay bribes abroad, the playing field is not levelled and unfair competition prevails. 136 Thus, on its own, peer review might not suffice to solve this collective action problem.
What are the theoretical implications of these findings? When it comes to reasons for noncompliance, the findings mainly correspond to an enforcement perspective, which asserts that states will comply with an international agreement when the benefits of compliance outweigh its costs. Interviews additionally suggest that low compliance activity is due to capacity and resource constraints, which corresponds to a management perspective. However, lack of capacity appears to be intertwined with lack of political will; how many personnel and resources states allocate to fighting foreign bribery is often (though not always) a political choice.
How about the policy interventions suggested by these theoretical schools? The policy solutions put forward by enforcement theorists, such as (soft) sanctioning in the form of peer and public pressure, help to raise awareness among investigators and prosecutors of foreign bribery, and create reputational costs. While these processes may have promoted some reforms in states, they are unlikely to tilt the balance towards full compliance on their own. The instances in which they appear most effective often coincide with the exposure of a corruption scandal or with the installation of a government that is more positively predisposed to fighting corruption. Once states are, in principle, willing to comply, the policy interventions suggested by the management school and constructivism (such as learning and international cooperation) become more important. These processes help to build capacity in states and foster skills and networks that are pivotal for detecting, investigating and sanctioning foreign bribery. The different peer review processes should therefore be seen as complementary; peer reviews might be most consequential when they can effectively deploy and combine these processes to address different barriers to state compliance. 137

Conclusion
Are peer reviews at all relevant for addressing transnational policy problems, such as foreign bribery? This article discussed three ways in which peer reviews can contribute to compliance with international anti-corruption norms: (1) by creating transparency about state behaviour; (2) by mobilising pressure on recusant states; and by (3) promoting learning and capacity building. Specifically, it analysed to what extent one such peer review (the OECD WGB) is perceived as capable of organising these processes, and how far participants in this peer review regard these processes as relevant for compliance.
To answer these questions, this article first developed a theoretical framework detailing the processes and mechanisms through which peer reviews were expected to contribute to compliance. This framework was inspired by three schools in the compliance literaturethe enforcement, management, and constructivist schoolsand existing research on peer reviews. Based on evidence from an online survey and semi-structured interviews, the empirical analysis shows that the WGB is indeed perceived as rather effective in creating transparency, organising pressure, and stimulating learning. Yet, the extent to which these processes can produce reform on the ground appears to be more limited and largely dependent on political will. 135 Jo-Anne Gilbert and J. C. Sharman, 'Turning a blind eye to bribery: Explaining failures to comply with the international anti-corruption regime', Political Studies, 46:1 (2014), pp. 74-89. 136 This not only applies to the WGB members that do not actively enforce the Convention, but also to several nonsignatory states with large export markets; interviews 1, 3, 8, 11, 14. 137 Delegates also mentioned how different peer review processes can be combined: for example, peer pressure and international cooperation (interview 1) and peer pressure and capacity building (interview 14). Enforcement scholars expect that states will comply with international norms, when the benefits of compliance outweigh its costs. The findings of this article largely confirm this expectation. From earlier research, we already knew that several governments are disinclined to comply with the Anti-Bribery Convention. 138 Similar considerations were raised in interviews conducted for this article: The perception that compliance harms business interests, the lack of interest in and awareness of foreign bribery in some states, and the efforts and resources that need to go into transnational bribery investigations, all raise the costs of compliance. Peer and public pressure are perceived to create awareness among law enforcement authorities of this crime and to impose material and social costs on states, such as having to redo an evaluation or negative media publicity. Even though these processes have triggered some change, these costs often do not outweigh the benefits of non-compliance. Furthermore, although many state delegates feel personally under pressure to take action, this pressure does not always trickle down to all responsible departments and actors.
While enforcement theorists offer the most compelling explanation for why states do (not) comply with anti-bribery legislation, this does not mean that we should refute the compliance strategies put forward by the management and constructivists schools. Political will might be crucial for compliance, but compliance is not only a matter of political will. Interviews indicate that states that are willing to fight foreign bribery benefit from the exchange of information as well as from the opportunities peer review offers for developing networks and personal relationships. Peer review processes aimed at promoting (social) learning, international networking, and capacity building are important in this regard. Sustained peer and public pressure, in turn, helps keep foreign bribery on the political agenda and gives much needed support to reform-oriented actors that are presently facing domestic resistance. This underscores that we should not assess peer reviews only in terms of their ability to trigger compliance, but also the extent to which they set in motion processes that contribute to state compliance in more indirect ways and over longer periods of time. 139 In light of the above findings, this article echoes earlier calls to treat the three compliance logics as complementary. 140 Several scholars have previously recognised that no compliance logic can account for states' behaviour on its own; in order to understand variation in compliance levels across states, we need to combine these three analytical perspectives. The findings of this article additionally suggest that reasons for (non-)compliance not only vary across states but might even differ within states over time. It is therefore important to consider how different compliance strategiesand by extension peer review processes 141can be combined to effectively promote compliance.
To conclude, what do these findings tell us about peer review as a mode of governance in the anti-corruption regime? As the oldest and most developed peer review in the anti-corruption regime, the WGB is a most-likely case to promote anti-corruption reforms. If the WGB faces challenges in promoting compliance, other peer reviews most likely do so as well. Of course, a single case study has its limitations. The WGB deals with fighting transnational corporate bribery. As argued earlier, this type of corruption is often not top of the political agenda and tends to garner limited public interest. In addition, the WGB brings together a fairly homogeneous group of industrialised states. Other peer reviews, which bring together a more diverse group of states, might face different or additional challenges when trying to promote compliance. Finally, this article only analysed the perceptions of individuals who participate in a peer review, many of which share responsibility for enforcing the Convention. They are, however, not the only 138 Gilbert and Sharman, 'Turning a blind eye to bribery'. 139 This corresponds to what Hale refers to as the 'compliance plus perspective' ('Under what conditions', p. 6). 140 Börzel et al., 'Obstinate and insufficient'; Carraro, 'Promoting compliance with human rights'; Hartlapp, 'On enforcement'; Tallberg 'Paths to compliance'; Versluis and Tarr, 'Improving compliance'. 141 Lehtonen also concludes that policy learning and accountability can be complementary in a peer review ('OECD Environmental Performance Review Programme'). responsible actors. An assessment of how political actors perceive compliance with the Convention, as well as their awareness of the WGB, provides an important avenue for future research.