Query

Please provide a synthesis of the literature on the effectiveness of integrity led interventions.

Introduction

Since the rise to prominence of corruption as a policy issue of global significance in the 1990s, there has been a core conceptual division with regards to appropriate responses to the problem.

On one hand, one set of policies has prioritised criminalisation, law enforcement and sanctions. These efforts have sought to ensure that corrupt behaviour is prohibited, increase the probability of detection and raise the costs of being penalised for corruption. On the other hand, a second set of measures has sought to advance a broader approach to preventing corruption through the promotion of integrity frameworks and ethical standards. While the former seeks to prevent corruption through exemplary punitive measures, the latter is focused on establishing positive examples and encouraging desirable behaviour.

Standards setting organisations, perhaps most notably the OECD, have been at the forefront of efforts on both fronts. This is exemplified by the near-contemporaneous publication of the

• 1997 Convention on Combating Bribery of Foreign Public Officials in International Business Transactions

and the

• 1998 Recommendation on Improving Ethical Conduct in the Public Service.

This two-pronged strategy is also reflected in the UN Convention against Corruption, where Chapter II promotes preventive measures such as conflict of interest policies, codes of conduct, transparent hiring criteria and integrity training, among others, while Chapter III deals with criminalisation. As such, it is clear that these efforts are meant to be broadly complementary, with “softer” tools intended to set norms and operationalise ethical standards being underpinned by the requisite legal tools to investigate and sanction potential instances of corruption.

As such, the view that the promotion of integrity frameworks and ethical standards can be a powerful antidote to poor governance and corruption is not novel. Indeed, the OECD’s 1998 Principles for Managing Ethics in Public Serviceset out many ideas that have since become core tenets of integrity led interventions (OECD 1998a; OECD 1998b). Among other things, The principles underscored the need to establish clear ethical standards with a legal basis; make these standards widely available to public servants; familiarise officials with their obligation to report wrongdoing; ensure transparency of decision-making; prescribe standards for interactions between the private and public sectors; emphasise the importance of ethical leadership by senior management; as well as codify ethical behaviour in operational policies and human resources procedures.

Since the late 1990s, numerous subsequent publications (including the OECD’s 2009 Towards a Sound Integrity Framework: Instruments, Processes, Structures and Conditions for Implementation) have provided a template for integrity focused approaches as a broad-based means of preventing corruption.

However, work by political scientists in the last decade indicates some intellectual tension over the appropriate prioritisation of direct, explicit anti-corruption measures vis a vis instruments that seek to reduce corruption indirectly by promoting alternatives to corrupt behaviour such as integrity (Heywood and Rose 2015; Doig 2012; Rose-Ackerman and Palifka 2016; Mulgan and Wanna 2011; Menzel 2015). Indeed, Heywood et al. (2017) lament that most anti-corruption interventions and policies have been designed with an overly narrow focus on ensuring compliance with and enforcement of legal standards. As Heywood et al. (2017: 3) put it, anti-corruption policies are often developed in response to certain scandals and typically target “institutional configurations or regulatory frameworks”.

Perhaps in response to the fact that policymakers have tended to view “corruption” as a policy problem that is best tackled through new legislation and tougher oversight, there have been growing calls for a renewed focus on the central role of values, ethics and integrity in controlling corruption (Heywood and Rose 2015; Menzel 2015). Influential organisations such as the OECD (2018a) and the World Bank (2015) have thrown their weight behind the campaign to highlight “the importance of accounting for behavioural elements in the formulation of anti-corruption and other development interventions”, as Camargo et al. (2020: 3) put it.

In parallel, a view has emerged in the corporate world that “anti-bribery and responsible business conduct compliance programmes […] can become legalistic, rules-based measures that do little to create a culture of integrity within companies” (OECD 2020: 80). Work by Langevoort (2016) concludes that such corporate compliance programmes are at best simply intended to avoid hefty fines and at worst simply cosmetic marketing tools (c.f. Krawiec 2003) and, as such, they have little impact on organisational cultures that tolerate integrity violations.

As a result of all of this, Meyer-Sahling and Mikkelsen (2020: 5) observe that, recognising that punitive measures are insufficient, policymakers and managers have increasingly “attempted to change ethics systems accordingly to include ‘softer’ ethics tools such as codes of ethics, appeals to ethical leadership, ethics workshops, and ethics training programs”.

The case for integrity centred approaches

Numerous scholars, drawing partly on insights from behavioural science, have in recent years interrogated the drivers of corruption. While the prospect of timely punitive measures for those found culpable of corruption are widely believed to affect individuals’ cost-benefit calculus, two core shortcomings of a legalistic or compliance led approach have been identified in the literature.

First, these approaches tend to rely on rational actor models that fail to satisfactorily account for norms or how group dynamics can affect individuals’ behaviour. Insights from social psychology studies based on real-world datasets have demonstrated that ethical considerations can outweigh financial incentives and cost-benefit analysis (see Tyler 2005). Moreover, individual morality is heavily shaped by social interactions with peer groups, as shown by a recent meta-analysis of 1,278 empirical studies investigating the psychology of morality (Ellemers et al. 2019).

Second, while stringent sanctions for corruption can shape behaviour, Wegner et al. (2013) argue that approaches based solely on penalising undesirable behaviour provide little inherent motivation for people to go above the minimum prescribed obligations. Indeed, citing literature from behavioural economics, Lambsdorff (2015) suggests that an excessive focus on oversight and penalties can “crowd out” individuals’ intrinsic motivation to behave ethically. Drawing chiefly on empirical and experimental studies on private sector anti-corruption compliance in the United States, he argues that coercive measures that deny people “the self-esteem of doing the right thing” can ultimately lead to diminished moral aspirations and ethical behaviour (Lambsdorff 2015: 4).

Lab based experiments into the drivers of behavioural change likewise suggest that approaches that appeal to individuals’ intrinsic motivation can lead to more consistent ethical behaviour by speaking to individuals’ morality and sense of responsibility (see Zúñiga 2018: 7-8). In this view, integrity frameworks that reward people for acting with and promoting integrity can generate incremental changes and enable the development of progressively more ambitious ethical standards. While acknowledging that a balance must be struck between “trusting the many intrinsically honest people and distrusting some corrupt”, Lambsdorff (2015: 5) claims that “this balance has shifted excessively towards distrust”.

Furthermore, even where anti-corruption measures are in fact able to reduce levels of corruption, Heywood et al. (2017) contend that integrity will not simply materialise in the vacuum left by reduced corruption. This, they argue is because integrity is not simply the inverse of corruption but a more expansive concept that “involves doing the right thing in the right way” (Heywood et al 2017: 3).

All this has led Zinnbauer (2019: 6) to conclude that:

In this view, legalistic compliance led approaches are ineffective, as one cannot specify rules for every eventuality, one cannot police all the rules all the time, and overregulation can present its own set of problems and indirect costs (Zinnbauer 2019: 4).

Despite such calls for a greater emphasis on integrity promotion, some academics point out that these efforts are less likely to be effective in high corruption contexts where progress in curbing corruption is most sorely needed. Mungiu-Pippidi (2017: 3), for instance, distinguishes between settings in which corruption is the exception, and those in which corruption is the norm, arguing that corruption is the “default governance order”. Where corruption is the norm rather than the exception, systemic pressures and incentives mean that Lambsdorff’s proposition, that the most effective anti-corruption measures are premised on the notion that only a minority people will behave in a corrupt fashion, is likely misguided. As Mungiu-Pippidi (2017: 3) puts it, “norm building and norm enforcement require two very different approaches”.

Strong theory, patchy evidence?

Despite the lively debate about the relative merits of “direct” and “indirect” approaches to curbing corruption, the evidence base remains patchy in terms of the actual effectiveness of the wide range of integrity instruments that have been prescribed as a means of reducing corruption.

Thus, a series of interrelated questions emerge. First, what evidence is there that integrity promotion instruments have actually been able to reduce corruption? Perhaps even more intriguingly, why – in the absence of clear evidence of their effect or impact – have certain integrity instruments been recommended by their proponents?

Second, how can the effectiveness of these integrity instruments be meaningfully assessed? Simply asking participants of integrity training sessions is unlikely to produce robust data of anything but short-term awareness. The use of proxy indicators is often complicated by their tenuous relationship between cause (e.g. code of conduct) and effect (lower reported incidence of corruption). Such knotty attribution problems are exacerbated by the fact that individual integrity instruments are rarely applied in isolation but typically in conjunction with other tools. More generally, one also encounters the common dilemma of how to ascertain and measure the impact of anti-corruption interventions (UNDP 2015; Wathne and Stephenson 2021; Heywood 2014).

Third, even where integrity instruments are believed to have some evidence of effectiveness, the experience of all public policy tells us that they are unlikely to be universally effective. Under which conditions and in which environments have specific integrity instruments produced encouraging results?

In an attempt to provide some answers to these questions, this paper presents a meta-analysis of existing literature. The remainder of this paper is structured by type of integrity instrument that has been proposed as a means of curbing corruption, from integrity training to codes of conduct. For each integrity instrument, studies that reveal something about its effectiveness are discussed.

This structured summary of the academic and policy literature aims to first marshal the available evidence in a manner that is helpful in assessing integrity measures thought to be effective in reducing corruption and in so doing inform policy and practice. Second, it attempts to serve as a “state of the evidence” and thereby inform investment in further research and rigorous testing of interventions.

Definitions

Before proceeding, it is instructive to provide a working definition of what is understood by “integrity” to determine how wide to cast our net, not least given the conceptual confusion that plagues the use of the term in different disciplines (Robinson et al 2018). As Heywood et al. (2017: 3) explain, “lack of clarity about what integrity is has hindered attempts to promote it”.

As can be seen from these definitions, the connotation of “integrity” varies in its meaning – from honesty to serving the public interest – and its scale– from individual morality to national cohesiveness. Zinnbauer (2019: 7-8) provides the following taxonomy of integrity led interventions:

• “Individual: to build personal integrity through ethics training, invocation of social value systems, and related awareness raising or priming techniques (c.f. Mazar et al. 2008; De Cremer et al. 2010).84c36ba40324
• Organisational: to nurture cultures of integrity within specific organisations, emphasising tone from the top, codes of conducts and an enabling intra-organisational ethics infrastructure (c.f. Kaptein 2008; Warren et al. 2014). [Kirby (2018) also proposes that organisational integrity encompasses the pursuit of an organisation’s “legitimate purpose” to the best of its ability.]
• Sectoral:bc33d1b5ca00 to build mutual trust and instil reciprocal commitment to integrity within a specific industry sector, and thus address the collective action dilemma that characterises many corruption situations and incentivises participants to deviate into corrupt behaviour as long as they expect their competitors to do the same (c.f. Persson et al. 2013). Such sectoral integrity initiatives include the extractives, construction, pharmaceutical, and shipping sectors (c.f. David-Barrett 2019)
• Systemic and integrated: to put in place holistic governance systems, either as national integrity systems or local integrity systems, that contain a web of mutually reinforcing transparency and accountability mechanisms to foster the integrity of all governance outcomes, although this approach is more geared towards structural than personal integrity (c.f. Six and Lawton 2013). Interventions designed to foster political integrity and promote democratic modes of government may also fall under this category.”

This literature review focuses primarily on integrity instruments at the individual and organisational levels. This is because theories of how anti-corruption interventions work typically adopt either the individual or the organisation as their unit of analysis (Meyer-Sahling and Mikkelsen 2020: 2). Consequently, most studies into the effectiveness of integrity tools likewise focus on the individual or organisational level. This is perhaps because, at higher levels of abstraction, such as the sector or even whole country, efforts to evaluate the effect of strategies and interventions lose coherence. For instance, determining the impact of ethical leadership in an entire sector such as the fisheries industry becomes extremely complex given the multitude of actors and jurisdictions involved.

While this paper does occasionally differentiate between tools primarily intended to enhance personal integrity from those seeking to promote an organisation-wide culture of integrity, it recognises that the relationship between these levels is often symbiotic. In other words, the behaviour of an individual is heavily influenced by the organisational culture in which they are embedded, while organisational integrity management frameworks rely on virtuous ethical behaviour by individual employees to succeed.

As Berry (2004) observes, where staff have a higher degree of personal integrity than the organisation, they are likely to become disillusioned and reluctant to report wrongdoing. Conversely, where employees’ personal integrity does not match the high standards set by the organisation, integrity failings may come to be viewed by staff as an inevitable response to “unrealistic” expectations. It is therefore paramount to ensure that members of an organisation “identify with the purpose of the organisation, know the rules and procedures, and understand how they are implemented in practice” (OECD 2020: 136-7). By the same token, it is important for organisations to display respect for personal integrity (namely, physical and mental integrity) as a means of reducing malpractices.

Integrity led interventions: An evidence review

The starting point for selecting interventions to consider in this paper is an OECD checklist for integrity management (OECD 2009). The list was then refined by way of reference to common approaches to promote integrity in the context of anti-corruption interventions.

Integrity training

Training forms a central part of the canon of integrity promotion and takes place primarily at the organisational level. Broadly speaking, such training is intended to (Meyer-Sahling et al. 2022: 4):

• “provide models of positive and negative behaviour
• provide guidance on rules, codes and methods and opportunities for practical application
• raise salience [awareness among employees] of ethical issues.”

Integrity and ethics training is often viewed as incorporating two strands: values and compliance. Value oriented training seeks to encourage staff to adopt a principled stance that can be referred to when they encounter ethical dilemmas. This type of training urges people to go beyond the minimum legal requirements to behave with honesty and integrity. Compliance oriented training instructs employees in their obligations and spells out penalties and rewards for different sets of behaviour to deter non-compliance and incentivise staff to report wrongdoing (Warren et al. 2014). While, in the past, these two approaches were seen almost as rivals (Whitton 2009), most modern integrity training encompasses both aspects.

Broadly speaking, training is a cross-cutting integrity instrument that is applied to convey values, knowledge and skills across a wide range of topics, which Resimić (2022: 3) notes can include:

1. ancillary activities (interests and activities of officials that could result in a conflict of interest)
2. anti-bribery and compliance
3. avoiding nepotism
4. codes of conduct and codes of ethics
5. dealing with freedom of information requests
6. income and asset disclosure
7. induction into an organisation, its rules, values and standards
8. managing interactions between the public and private sectors (conflict of interest provisions, lobbying, pre- and post-employment rules)
9. receipt of gifts and hospitality
10. responsible business conduct and corporate social responsibility
11. whistleblowing mechanisms

Integrity training is widely held to be a valuable tool to strengthen awareness about corruption prevention (OECD 2013: 9). Cochrane (2019: 10) even points out that the “unquestioned assumption that integrity education will reduce levels of misconduct and corruption is long-held and pervasive”. By raising awareness of ethical topics and integrity obligations, the theory goes that staff acquire knowledge, revise existing attitudes and ultimately change their intentions and actual behaviour (Meyer-Sahling et al. 2022).

Yet, while anti-corruption and integrity training is by now commonplace, aggregate evidence about its impact is elusive. An OECD stocktaking exercising found that, although the majority of OECD countries evaluate the quality of all training activities, they do not tend to measure their impact (Pearson 2011). Meyer-Sahling et al. (2022: 3) note that “beyond observational studies in [the] private sector and with business school students”, evidence is scarce.

Part of the problem is methodological – there is uncertainty about how to measure the effect of training in changing individual behaviour and organisational culture (Van Montfort et al. 2014). Indeed, such training is commonly evaluated using activity level indicators, such as the number of employees or officials trained or satisfaction surveys that reveal more about participants’ enjoyment of an event than its learning outcomes (Cochrane 2019). In Hungary, for instance, Pallai and Gregor (2016) found that participant satisfaction was very high, while learning assessments suggested there had been “minimal to zero learning impact”. As such, integrity training can assume characteristics of a tick-box compliance instrument, whereby all staff are obliged to undergo training, but little effort is made to ascertain impact on behaviour.

Moreover, studies on the effectiveness of integrity training have been hamstrung by the fact that information about the values and behaviour of staff is often considered sensitive and thus can be difficult to collect, especially over time as part of longitudinal panel studies (Treviño et al. 2006).

The numerous empirical studies (Kaptein 2009; Kaptein 2011; McKendall et al. 2002; Treviño et al. 1998; Treviño and Weaver 2001; Weaver and Treviño 1999) that have attempted to evaluate the impact of integrity or ethics training on organisational outcomes using cross-sectional data are ill-suited to inferring causal effects. This is because these studies only observe the sample population at one point in time; they are unable to indicate whether a change in one variable can generate a change in another variable over time. There may be a natural correlation between those organisations that are generally more ethically oriented and those that require their employees to undergo rigorous integrity training, suggesting that “ethics programs may not be the cause, but the outcome, of more ethical attitudes and behaviour” on the part of organisations (Warren et al 2014: 86).

The OECD (2020: 127) also points to the attribution problem that arises due to the “many variables that can influence the integrity of public officials who participated in training activities [which] make measuring the outcomes of training complex”. Moreover, given the highly variable content of integrity training, it is difficult to draw sweeping conclusions about their effectiveness.

Training can take different forms, from e-learning modules to seminars and dilemma training (see below). Participation might be obligatory or mandatory, can be provided by an in-house team or external training providers, and the extent and nature of follow-up on the training session varies considerably (Van Montfort et al. 2014). While there is a widespread assumption that the content and format of integrity training matter in terms of its impact, there has been little comparative empirical work done into the factors that influence an integrity training session’s effectiveness (Van Montfort et al. 2014).

Behavioural science provides some useful insights into what types of public integrity training are most effective didactically (Whitton 2009). Studies suggest that integrity training should combine learning rules and principles as well as building knowledge and problem solving skills (Jackson and Köbis 2018: 36). There is some suggestion that dilemma training may be more effective than less interactive training formats. In dilemma training, participants are faced with real-world situations, including challenges they have encountered in the past, and are supported to plot out appropriate ethical responses. Bazerman and Tenbrunsel (2011) note that such sessions offer learning opportunities and practice guidance based on real-world scenarios and as such may be taken more seriously by staff than other training formats. For its part, the OECD (2020: 126) asserts that basing dilemma training on realistic situations “helps stimulate participants’ moral awareness, contributes to their level of moral reasoning, and provides methods to help improve the moral quality of their actions”. However, it does not provide empirical backing for these claims.

Social psychologists have documented how face-to-face communication leads to improved trust and better learning outcomes (Drolet and Morris 2000). Consequentially, in-person integrity training with small groups is considered more effective than virtual online sessions, not least as they lend themselves to realistic case scenarios (Warren et al. 2014).

Overall, the evidence on whether training can be effective in building resilience and commitment to integrity is mixed. On one hand, multiple studies of the private sector have concluded that integrity training has no discernible outcome on ethics outcomes (Kaptein 2011; McKendall et al. 2002; O’Fallon and Butterfield 2005; Delaney and Sockell 1992). Similar conclusions have been drawn in the public sector (Menzel 1997: 224; Van Montfoort et al. 2014; West and Berman 2004: 189).

Other more recent studies of integrity training in both the private and public sector have produced somewhat more encouraging results (Warren et al. 2014; Pallai and Gregor 2016; Meyer-Sahling et al. 2022). Warren et al. (2014) found that the introduction of formal ethics training in US banks was associated with less observed unethical behaviour among staff two years after the training, while a study by Meyer-Sahling et al. (2022) concluded that ethical leadership training of police officers in Bangladesh increased officers’ willingness to decline gifts. Conceivably, this might indicate that the design and delivery of integrity training is improving over time.

On balance, it appears that training is a necessary but far from sufficient condition to promote organisational integrity. Its main value is to familiarise staff with organisational values, rules, procedures and policies so that none can claim ignorance. Studies in social psychology have documented the key role of “common knowledge” as a prerequisite for ensuring effective coordination, such as that required to operate an integrity framework: it is important not only that standards are widely disseminated but also that everyone knows that everyone else is familiar with these standards (Thomas et al. 2014). As such, integrity training is thought to be of particular importance to new members of a given organisation as a communication tool to set the tone for expected standards of behaviour, and to stress shared organisational values among staff (OECD 2013).

To enhance effectiveness, the OECD (2018a: 40) suggests it should be periodically repeated and updated, and training can perhaps be best seen as a delivery mechanism to raise awareness to the existence and function of other integrity instruments. Lessons from the use of gender training to mainstream gender equality in organisations indicates that one-off training might generate superficial awareness but shifting behaviour in a meaningful way requires longer term engagement and repetition (Callerstig 2016: 119–120).

Polaine (2018: 45) recommends embedding integrity training into broader methods to ensure professionalism, such as by including active participation in such training as part of an employee’s performance assessment. As such, it is clear that, despite the mixed evidence of effectiveness, there is little suggestion that integrity training should be abandoned altogether. Rather, the question is how to design training content and configurations to support wider organisational integrity management frameworks (OECD 2013). One key lesson appears to be that, despite the proliferation of online training opportunities, face-to-face sessions that cover both values and compliance aspects and are based on experiential learning methods are most likely to be effective (Warren et al. 2014). Nonetheless, there is a continued need for more refined methods to assess attitudinal and behaviour change that might arise as a result of integrity training.

In this regard, Cochrane (2019: 4) provides a useful overview of potential approaches to evaluate the impact of integrity training (see Table 2).

Ethical leadership

People take their social cues from authority figures. In both public and private organisations, the ethical behaviour of managers is believed to be one of the main channels to promote integrity among staff and raise employees’ moral awareness (OECD 2020: 73). The visible commitment of leaders to high ethical standards is thought to contribute to a culture in which employees feel comfortable speaking up and reporting wrongdoing (Avey et al. 2012; Detert and Burris 2007; Walumbwa and Schaubroeck 2009). Theoretical models have long suggested that leaders who encourage open organisational cultures are able to reap the rewards associated with having employees who feel engaged and valued (Beugré 2010), findings corroborated by a recent empirical study based on an online survey of 1,039 Australian nurses (Holland et al. 2017). Ete et al. (2022) even propose that integrity has become “an axiom for leadership effectiveness”.

Empirical research suggests that ethical leadership is important for the integrity of public administration. A longitudinal study by Beeri et al. (2013) of 108 employees in an Israeli regional council found that ethical leadership was positively related to employees’ awareness of the code of ethics. Surveys conducted by Hassan et al. (2014: 1) in a large agency in state government in the US also indicated that “after controlling for the effects of employee characteristics, perceptions of procedural fairness, and supportive leader behaviour, ethical leadership reduced absenteeism and had a positive influence on organisational commitment and willingness to report ethical problems”.

In the anti-corruption world, there is a widespread consensus that “tone from the top” is critical in both the public and private sectors (New South Wales Independent Commission Against Corruption 2019; Dent 2021; The American Anti-Corruption Institute 2017; GAN Integrity 2021). This applies to senior figures within a given organisation, such as managers or executives, as well as to political leaders. Lambsdorff (2015: 10) has even asserted that the tone from the top could be “the most important factor in fighting corruption”. A recent study of corruption in the health sector in Rwanda, Uganda and Tanzania found that (Camargo et al. 2020: 10):

It is reported that in Tanzania, public officials began simply stating the name of the country’s then president Magufuli to indicate that it was not possible for them to accept bribes, based on the perception that the president was “a credible anti-corruption crusader” (Camargo et al. 2020: 10). This is suggestive of the ways in which individuals use leaders as ethical reference points when making decisions.

Conversely, where leaders openly flout rules, procedures and ethical standards, integrity frameworks are likely to become dysfunctional very quickly. A study by Hanna et al (2013) into 115 higher education institutions across 36 different countries concluded that one of the most significant drivers of unethical behaviour among staff in organisations is the observation of managers engaging in wrongdoing as more junior staff “learn” from their leadership and imitate their behaviour.

The fact that a perception of ethical leadership derives from largely intangible qualities such as leaders’ “demeanour, attitude and reputation” makes it “difficult to prescribe how to achieve it or test for it” (New South Wales Independent Commission Against Corruption 2019). Moreover, as Lambsdorff (2015: 11) notes, “the tone at the top is not objectively measurable”.

Nonetheless, certain tools are thought to be useful for leaders to demonstrate their commitment to integrity. At the organisational level, these tools range from regular bilateral meetings between manager and employee to the development of recruitment and performance management systems that include integrity as an assessment criterion (Tangirala and Ramanujam 2012). Overall, the OECD (2020: 97) suggests that the most effective means of setting this tone and inculcating organisational integrity is “the day-to-day work environment and the daily interactions between leader and follower”.

Beyond managers simply signalling commitment to ethical values, which could risk being seen as tokenistic, there is increasing scholarly interest in tangible ways to strengthen leaders’ personal integrity. The literature indicates that leaders’ integrity is a function of their cultural and environmental influences, relationships and lived experiences (McKenna and Campbell 2011; Seijts et al. 2015; Sosik et al. 2012). Logically, therefore, it follows that personal integrity is malleable; it can be shaped through targeted interventions such as training and mentoring. In particular, leadership coaching is increasingly viewed as a useful means of influencing people’s behaviour and performance (Anthony 2017).

A recent study explored the potential impact of coaching on managers’ personal integrity in corporate South Africa. Based on six semi-structured interviews, Van der Walt and Van Coller-Peter (2020) contend that a small number of coaching sessions can contribute to leaders’ awareness to the importance of integrity. In turn, greater awareness of integrity was reportedly perceived by study participants as having enhanced their ability to act in “accordance with stated values” and use “ethical considerations to guide decisions and action” (Van der Walt and Van Coller-Peter 2020: 2). The authors explain this by suggesting that the coaching sessions offered leaders a “safe space for meaningful conversations… and become immersed in issues around integrity” (Van der Walt and Van Coller-Peter 2020: 6). In the view of the authors, the findings demonstrate that coaching is a suitable tool to raise awareness to integrity among leaders. However, given the small sample size and the fact that the researcher was a coach employed at the same organisation as the study participants, the external validity of the findings remains open to question.

Peer group role models

Individuals’ behaviour is shaped not only by the actions of their leaders but also those of their wider peer group. Where other individuals commit wrongdoing, their peers tend to become more tolerant of such behaviour, especially in cases where it is not punished (Gächter and Schulz 2016). As such, behavioural techniques to challenge norms that tolerate corruption are increasingly recognised as a potential avenue to enhance integrity (Nicaise 2021; Jackson and Köbis 2018; Kassa and Camargo2017).

Clearly, norm-building strategies are not a short-term project, but Muers (2018: 12) contends that deliberate government policies can play an important value-signalling function and contribute to incremental cultural changes over time. Collier (2016) asserts that integrity instruments that reward integrity and censure corruption can help to increase awareness about the gap between behaviour that was previously tolerated and that which is now acceptable.

One example of such a signalling mechanism is integrity awards, which are employed by public administrations03d5c69f3ef4 and conferred by civil society groups (Adams 2020; Integrity Icon Sri Lanka 2021). One of the best known such recognition oriented initiatives is Integrity Idol, recently rebranded to Integrity Icon. Pioneered by the Accountability Lab, Integrity Icon is a global campaign launched in 2014 by the Accountability Lab explicitly on the premise that “direct” anti-corruption approaches (Accountability Lab 2021):

The campaign now operates in 12 countries to “name and fame honest government officials, change the narrative around graft and rebuild trust in government through lifting up role-models” (Accountability Lab 2021). However, despite the claim that “all of the evidence indicates that positive reinforcement is what changes behaviour” (Accountability Lab 2021), the precise impact of recognition led approaches to building integrity remains contested.

On one hand, it is clear that Integrity Icon has been successful in reaching a large number of people in the countries of intervention. By 2017, there were 4,689 nominations for the Integrity Idol in Liberia, and the campaign reached an estimated 4 million people through television and radio (Wijesinha 2018). The Accountability Lab (2021) also points to some notable results, including the appointment of a competition winner as the minister of justice (reportedly as a result of winning the award), and the role given to another winner on a government ethics committee.

Based on qualitative interviews with nominated idols and their colleagues in Mali, an MIT researcher found evidence of widespread acknowledgement that the Integrity Idol project had produced a potent media campaign and brought attention to integrity issues in local communities (Accountability Lab 2018). However, in terms of diffusing norms of integrity, the impact was reportedly mixed. On one hand, in some institutions, winners’ colleagues began to emulate their behaviour. On the other hand, in hierarchical institutions marked by little interpersonal trust, such as the Malian army, interviews revealed resentment on the part of superior officers towards the publicity received by the idols. In one case, senior officers even transferred a solider nominated as an integrity idol to a more difficult posting (Accountability Lab 2018).

Thus, in systemically corrupt settings characterised by mistrust, it appears that “faming” can occasionally backfire. Where non-financial incentives such as recognition are not accompanied by structural reform and supported by ethical leadership, raising the profile of high integrity individuals can actually lead to negative outcomes for them. If the intervention logic of Integrity Icon holds true, this could in fact discourage icons’ peers from behaving in a more ethical manner.

More robust evidence on the utility of this “recognition led” approach has recently emerged from a randomised field experiment and accompanying ethnographic field study in Uganda. The setting was a national park in which wildlife conservation efforts and revenue for local people is generated through tourism. Buntaine et al. (2022a) sought to evaluate the theory that providing positive recognition “for local leaders who forego corruption” in the revenue sharing scheme in Uganda would lead to lower levels of corruption. It was hypothesised that providing recognition to local committees for effective management of a revenue sharing project through radio and award ceremonies would foster collective pride and reduce malfeasance. Results from villages exposed to the campaign were compared with a control group of villages that did not participate in the study.

Contrary to the intervention logic of Integrity Icon campaigns, the authors found that “the offer of recognition did not change leaders’ behaviours or attitudes about corruption and governance. Nor did the offer of recognition result in better project outcomes” (Buntaine et al. 2022a: 2). Likewise, campaigns that informed citizens about the existence of the awards did not result in changes to public attitudes about corruption.

Given that neither intervention improved outcomes or changed anti-corruption norms, Buntaine et al. (2022) concluded that symbolic recognition of integrity alone is insufficient to reset norms, expectations or behaviours. Particularly notable was the inability of recognition alone to motivate people; multiple subjects in the field study expressed a clear preference for tangible rewards, such as money or goats in exchange for acting with integrity. This speaks to a more pragmatic and transactional view of the benefits of integrity, a sense that is reinforced by the finding that committee members felt that the incentives were overly focused on rewarding whole communities rather than them as individuals (Buntaine et al. 2022b: 18).

The evidence from Uganda indicates that recognition centred approaches “will not easily alter behaviours and nor are they likely to change more fundamental norms that operate within public life” (Buntaine et al. 2022b: 2). As such, there is some suggestion that positive recognition campaigns are best viewed as an adjunct to more system oriented approaches that foreground structural and material concerns, including salaries, promotions and elections. While these findings suggest that recognition oriented interventions are ineffective in isolation, they at least do not appear to have “harmed anti-corruption efforts” in Uganda (Buntaine et al. 2022: 3). Given their relatively low cost, the authors recommend “adding recognition as an additional component” to other anti-corruption measures (Buntaine et al. 2022b: 3).

To the extent that positive recognition can complement and reinforce reforms that target instrumental concerns, it seems likely that these virtue-signalling interventions will need to be consistently applied over long periods of time. In the view of Buntaine et al. (2022b), recognition and reward should take place at multiple levels of government and combine symbolic and material elements. Adams (2020) points to two other lessons with regards to integrity awards. Where such competitions are organised by governments, then can themselves become an avenue of patronage and nepotism and thus backfire. Moreover, public participation in the selection of integrity idols is important to the credibility and sustainability of these initiatives.

Addressing social norms through ‘nudges’

The growing interest of some scholars of corruption in behavioural sciences in recent years has resulted in increased attention being devoted to studying so-called nudges. In the words of Camargo and Burgess (2022: 30),

Nudging has been proposed as an effective tool to change undesirable behaviours that have been inadequately tackled by traditional policy interventions, like education campaigns, training or financial incentives (Thaler and Sunstein 2008). The focus is on tweaking the environment in which people make choices to guide them towards the desired outcomes without directly restricting their freedom of choice. Typically, these measures target specific groups – such as public officials working in procurement – with crafted messages that draw on behavioural insights such as the attitudinal drivers of corruption. Findings from other fields, such as tax collection, seem encouraging. Hallsworth et al. (2017) conducted a natural experiment involving 200,000 individuals in the UK, which found that targeting taxpayers with specific social norms and public service messages improved tax collection.

Moreover, meta-analysis of the field indicates that nudging campaigns are often successful. In their review of 100 empirical articles evaluating the effect of nudging in a range of different sectors, Hummel and Maedche (2019) reported that 62% of interventions led to a change in behaviour that was statistically significant. There is, however, an important caveat; the size of the effect is generally small, indicating that not everyone is responsive to these kinds of interventions.

The evidence indicates that nudges generally affect people who do not have strong pre-existing preferences, while individuals with ingrained habits and attitudes are thought to be fairly resistant to these measures. Indeed, according to Venema and van Gestel (2021: 224), nudges are best suited to settings in which “people are indifferent to the behaviour at hand, when they have good intentions that they forget about, when they experience conflicting preferences, and in novel choice contexts where people do not know what to do”.

Few of these conditions ring true in the area of corruption, where some people have strong preferences in favour of corruption (perpetrators, like bribe-takers) while those who have preferences against corruption (victims, like bribe-givers) are generally less powerful or stuck in a collective action trap. Consequentially, corruption, as an entrenched set of behaviours that materially benefit powerful insiders may not be the most promising domain for nudge-based interventions.

Despite these reservations, lab experiments with students conducted by Köbis et al (2015) indicate that small normative prompts can affect the decisions of individuals confronted by corrupt scenarios. However, although “behavioural interventions have proven effective in experimental settings”, there is limited real-world evidence on how behavioural interventions can “motivate decision-making that supports better control of corruption” (Kassa and Camargo 2017: 3). In addition, understanding how policy makers can operationalise the behavioural insights as part of anti-corruption efforts is complicated by the fact that “there are no blueprints to predicting what nudges people will respond to and what way” (Kassa and Camargo 2017: 3).

The literature suggests that the content, tone and timing of informational nudges all influence their effectiveness. Merely providing people with information about their integrity obligations does not mean that they will pay heed to it. The OECD (2020: 122) argues that moral reminders have the potential to positively influence individuals’ behaviour, but that this information is most effective when provided at the moment of decision-making (c.f. Mazar and Ariely 2006; Pruckner and Sausgruber 2014).

A recent experimental study based on a sample of 5,000 public officials in Chile nonetheless supported the view that normative prompts can help nudge people towards greater integrity. Meyer-Sahling et al. (2019) used a survey instrument to ascertain the effect of “activating” public service motivation. Providing civil servants with prompts emphasising public service values was found to increase their reported willingness to report ethical problems to management. Meyer-Sahling et al. (2019: 11) therefore contend that activating a sense of public service values is both “feasible through low-intensity treatments” such as prompts and “beneficial for public sector ethics”.

The policy implication here is that providing public servants with reminders of their ethical obligations at critical decision-making junctures can enhance public sector integrity. There are several important caveats. The effects were minimal among respondents whose level of “public service motivation” was assessed to be low; this implies that staff with a low level of personal integrity are not likely to respond positively to this kind of social nudges. Second, the authors note that their findings may only apply to countries with low levels of public sector corruption, in other words where corruption is the exception rather than the norm. Finally, the prompts were delivered immediately prior to respondents stating whether they would be willing to report wrongdoing to management. Stated intent is not the same as actual behaviour. On top of this, the Meyer-Sahling et al. (2019) note that they do not know how long-lived these social nudges are on public officials’ ethical attitudes. The limited longevity of the impact of nudge messaging on the target population has also been noted elsewhere (Hallsworth et al. 2017).

Falisse and Leszczynska (2022) conducted a “lab-in-the-field” experiment with public officials in Burundi to assess the impact of informational anti-corruption nudges on behaviour. Different groups of officials were exposed to messages that either appealed to (1) the general idea of good governance or (2) their professional identity before being asked to allocate a rationed public good among citizens, some of whom offered bribes. The study found that officials who received messages appealing to their professional identity were found to be more likely to distribute goods equally between citizens than officials who were not exposed to any message.

However, neither type of message affected officials’ propensity to accept bribes, or the size of the bribe accepted. The broader good governance messaging also had no discernible effect on officials’ tendency to allocate goods fairly. The authors conclude that nudges that seek to appeal to someone’s self-image, such as a sense of professional duty or identity, are more effective than generic anti-corruption messages because they make it more difficult for people to “disconnect their actions from their identity”. However, even the limited impact of appeals to professional identity were only found to affect “less experienced participants” (Falisse and Leszczynska 2022). This is in line with the findings by Venema and van Gestel (2021: 224) that nudges only affect people who are not already set in their ways.

Other recent research into how behavioural factors such as social pressures and shared beliefs sustain petty corruption in East Africa demonstrates that behavioural drivers are intrinsically linked to systematic pressures, such as the dearth of accessible healthcare. In these settings, Camargo et al. (2020) argue that corruption effectively serves as a means to solve problems such as the lack of access to quality public services. The authors find that, in highly corrupt settings, even individuals that disapprove of corruption are likely to engage in it because social pressures outweigh their individual values, suggesting that interventions aimed at bolstering personal integrity will be ineffective where they fail to account for systemic pressures. This is supported by Falisse and Leszczynska (2022) who note that in high corruption settings like Burundi, the impact of interventions such as integrity messaging intended to act as a “soft reminder of social norms and the social costs associated with favouring the briber” is likely to remain limited where it does not address the material drivers of corruption, such as very low public sector salaries.

The OECD (2022: 13) likewise cautions against “a false understanding of behavioural science [that] may lead some to believe a structural problem is a behavioural one”, and suggests deploying behavioural insights “to enhance rather than substitute more classic policy-making”.

Therefore, due consideration of social and behavioural factors should not neglect the core material drivers of corrupt behaviour. Otherwise, despite the best efforts of behavioural focused campaigns, norm nudges or integrity promotion initiatives, people will continue to rely on social networks in a nepotistic and corrupt fashion. Camargo et al. (2020: 13) therefore contend that behavioural nudges are most effective when deployed in conjunction with “strong enforcement of sanctions”. Importantly, they argue that, as demonstrated in Rwanda and to a lesser extent in Tanzania, “top-down enforcement of the rule of law plays a significant role in transforming expectations and in shaping the willingness of individuals to partake in illegal actions”.

Moreover, unlike other scholars’ insistence on the need for “positive” messaging and reinforcement, Camargo et al. (2020) propose that shaming can also be effective in shifting expectations and behaviours in the desired direction. For instance, in Rwanda, the authorities “naming and shaming” approach means that being caught for corruption brings disgrace on an individual’s entire social network. This wider normative social cost helps to narrow the “gap between formal and informal rules” (Camargo et al. 2020: 12). Similarly, in Indonesia, a fatwa issued in 2014 against the illegal wildlife trade demonstrated how top-down interventions can build on local deeply rooted social norms (Wallenand Daut 2018).

In sum, Camargo et al. (2020: 13) argue that, while behavioural approaches are not an answer in their own right, “they can usefully complement conventional principal-agent based reforms to increase the perceived costs of engaging in corrupt actions”. The implication is that the most effective strategies are likely to be two-pronged: seeking to alleviate systematic material drivers of corruption (such as scarcity of public goods) while also incorporating norm-focused interventions that seek to “change public perceptions about the inevitability of corruption” (Camargo et al. 2022: 14).

Another study, by Camargo and Burgess (2022: 33), of the behavioural drivers of corruption that facilitate the illegal wildlife trade documents that some of the most effective efforts deployed behavioural insights “to complement conventional programmes”. For instance, in China and Vietnam a two-track model was deployed. One component revolved around communicating a robust law enforcement response to ensure that the law was perceived as an effective deterrent, while also looking to nudge consumer behaviour to reduce demand for illegal wildlife goods. In this case, the nudge involved targeted messaging aimed at the chief consumers of rhino horn (middle-aged businessmen) that “success, masculinity, and good fortune come from an individual’s strength of character and not from the use of rhino horn” (Camargo and Burgess 2022: 33).

As such, Camargo and Burgess (2022: 43-44) stress the importance of robust diagnosis of the underlying drivers of corrupt behaviour, as “some drivers will be amenable to a behavioural approach, in other [settings] a behavioural element can usefully complement other approaches, and sometimes a behavioural approach will be inadequate”.

Likewise, the OECD (2022:11) has stated that:

Indeed, even advocates of nudge-based approaches acknowledge that (Köbis et al 2019: 20):

Anti-corruption awareness raising campaigns

Alongside social nudges and other targeted measures that seek to promote the virtues of acting with integrity, broad-based awareness raising or public information campaigns are also potential vectors to encourage integrity. The theory is that such campaigns can create a “conscious link between an individual’s view of his or her own integrity and the wider public benefit” (OECD 2020: 85), and make sure that people do not perceive corruption as an abstract, victimless crime (c.f. Barkan et al. 2015). Here, media coverage is thought to be helpful to diffuse norms about socially acceptable behaviour and challenge collectively held “mental models” that facilitate corruption (Camargo and Burgess 2022: 30). The idea is that visibility campaigns that deploy anti-corruption messages on billboards, posters, television, community radio, social media and so on will encourage people to refuse to engage in corruption and report wrongdoing (Peiffer and Alvarez 2016). In line with the importance of tone from the top and ethical leadership, Camargo and Burgess (2022: 29) argue “it matters who delivers the message” and “influential personalities, celebrities, religious leaders, and important business or political figures” can help accelerate the dissemination of new norms.

However, while awareness raising campaigns are “one of the least expensive tools of the anti-corruption arsenal”, for a long time their effectiveness remained poorly documented (Falisse and Leszczynska 2022). In recent years, a number of studies have sought to interrogate the impact of these types of interventions and have yielded some surprising insights and unintended consequences.

Kassa and Camargo (2017: 3) contend that the “dissemination of strong and carefully tailored messages may act as a catalyst to trigger changes in public attitudes instilling an expectation that corruption can indeed be curbed… there is growing evidence that ‘edutainment’ campaigns can be effective.” Camargo et al. (2020: 9) point to the Rwandan government’s civic education programmes as a successful example of how to strengthen “shared mental models that promote positive control of corruption”.

At the same time, evaluations of social norms interventions in the health sector have concluded that “publicising the prevalence of a harmful practice can make things worse” (Cislaghi and Heise 2018). There is some compelling evidence that this may also be the case in the area of anti-corruption. Peiffer’s quantitative research in Indonesia, Nigeria and Papua New Guinea suggests that raising people’s awareness about the extent of corruption can actually reinforce the sense that it is intractable and therefore reduce citizen’s willingness to report wrongdoing or support anti-corruption efforts (Peiffer 2017; Peiffer 2018). This echoes earlier findings by Chong et al. (2015) that awareness raising campaigns in Mexico reaffirmed citizens’ cynicism about the prevalence of corruption; providing citizens with information about the level of mayoral venality did not affect the extent to which they thought the municipal government was corrupt.

Cheeseman and Peiffer (2020) suggest that increased awareness about corruption can exacerbate the collective action problem that plagues anti-corruption efforts. Even carefully crafted, positive messages designed to encourage citizen action against corruption – such as the ease with which people can report malfeasance – have been found to backfire and result in reduced appetite to oppose corruption (Peiffer 2017). Particularly in societies in transition, messages that reinforce the sense that corruption is widespread contribute to a sense that it is the norm rather than the exception, and can therefore even encourage people to behave more corruptly (Cheeseman and Peiffer 2020: 10).

While previous studies assessed the impact of anti-corruption messaging on people’s attitudes towards corruption or their reported willingness to bribe, Cheeseman and Peiffer (2020) focused on how these messages influence actual behaviour in Nigeria. The authors found that anti-corruption messaging had either no effect or actually resulted in those exposed to the messages becoming more likely to pay bribes. Cheeseman and Peiffer (2020: 4) explain the results by suggesting that:

The negative impact of anti-corruption messaging was particularly pronounced among people who already believe corruption is pervasive. Overall, Cheeseman and Peiffer (2020: 4) conclude that “untargeted anti-corruption messaging is not just a waste of money but may actually make it harder for other strategies to succeed”.

Other recent studies that have considered the nuanced application of anti-corruption messaging campaigns have produced some more encouraging findings. In Papua New Guinea, Peiffer and Walton (2019) found that, contrary to the findings in Indonesia, exposure to anti-corruption messages did not trigger pessimistic views about reporting corruption or an unwillingness to report corruption. Three different types of messages were trialled, those that were morally themed, those that stressed the illegal nature of corruption and those that emphasised corruption’s harmful impact on local community and kinship groups. While the “moral” and “legal” messages had limited impact, messages related to the communal effect of corruption enhanced citizen’s reported willingness to support anti-corruption efforts. Peiffer and Walton (2019: 25) therefore suggest that not all awareness raising efforts result in corruption fatigue, and “the right anti-corruption message can improve citizens’ chances of reporting corruption”. While messages about the extent or illegality of corruption proved to be “unable to motivate many [people] to think and act differently when they observe corruption occurring or otherwise confront it”, messages related to the social cost on the local community appear to be more promising (Peiffer and Walton 2019: 26).

Finally, Köbis et al. (2019) insist on the importance of distinguishing between anti-corruption awareness raising campaigns that focus on “injunctive norms” from those that deploy “descriptive norms”. While injunctive norms relate to what is deemed socially or morally acceptable, descriptive norms refer to what is believed to be common forms of behaviour. In other words, an injunctive norm implies that people engage (or not) in a type of behaviour because they believe others expect this from them, whereas a descriptive norm implies that people engage in certain behaviour because they believe this to be widespread.

In their lab-in-the-field experiment in South Africa, they studied the effects of a poster campaign that tried to affect people’s perceptions about the behaviour of other people, instead of trying to raise citizen awareness about the negative impact of corruption. The messages therefore relayed dynamic trends (“Less and less South Africans bribe”) than static information (“12 percent of South Africans bribe”). By displaying positive descriptive norms about the behaviour of fellow citizens, Köbis et al. (2019) suggest the campaign was able to avoid the backfire effect of awareness raising campaigns observed by Peiffer in Indonesia.

The authors found that, after exposure to posters conveying this kind of messaging, participants perceived corruption to be less common, and were less likely to engage in bribery in a mobile lab experiment involving real money. They suggest that targeting descriptive norms about the perceived incidence of corruption is more important than trying to convey the idea that corruption is unethical, because “when descriptive and injunctive norms are incongruent, descriptive norms often exert a stronger pull”. In other words, “people often bribe not because they consider it the right thing to do, but rather because they have few other options due to the perceived high frequency of bribery around them” (Köbis et al. 2019: 19-20).

Köbis et al (2019: 19) contend that their lab-in-the-field approach addressed some of the concerns about the external validity of lab experiments, and provided some “indications that norms trend messages outside of the lab can reduce both perceptions and behaviours of corruption inside the mobile lab”. However, as is often seemingly the case with behavioural campaigns, the study also found that perceived social norms and proclivity to engage in bribery swiftly returned to pre-treatment levels.

Overall, the lessons learned in the field of anti-corruption messaging seem to imply that instead of reinforcing what citizens already perceive to be the case – that corruption is rampant and immoral – these types of interventions need to challenge “conventional wisdoms” about the expected (corrupt) behaviour of fellow citizens (Camargo et al. 2020). By challenging descriptive norms about the perceived intractability of corruption, awareness raising campaigns may help societies escape what Stephenson (2020) refers to as the “self-reinforcing trap”, where people engage in corruption because they expect others to behave in a corrupt manner.

Ombuds offices

Ombuds offices, which originate from 19th century Sweden, have grown in popularity since the 1960s and are now present in many countries around the world (OECD 2018b: 4). As oversight bodies tasked with processing citizen grievances and providing general awareness and counselling, they contribute to integrity management in different ways. However, assessing their impact in improving integrity and reducing corruption is complicated by three factors.

First, curbing corruption has not traditionally been at the forefront of the ombudsman’s tasks, which typically focus more broadly on “addressing grievances related to administrative issues” (International Ombudsman Institute 2022: 3). Second, even as they have come to be recognised as having a role to play in curbing corruption, their exact mandate, role and relationship to other institutions and the justice system continues to vary by country, making a comparative impact analysis difficult (Council of Europe 2021). Finally, there is little consensus on the metrics according to which their effectiveness could be assessed. Stuhmcke (2014: 2) observes that standard cost-effectiveness measures are unable to capture the “intangibility” of ombuds offices’ contribution to the public good and their role in promoting values like “integrity, fairness, equity and humane treatment”.

Evaluations have also had to grapple with an attribution problem as ombuds offices are part of a wider integrity system and tend to have little coercive power over other government agencies (OECD 2018b). This means ascertaining the precise nature of their contribution to any reduction in the level of corruption or increase in citizen trust is problematic. More generally, indicators that have been used as proxies for the effectiveness of ombuds offices are imperfect. A reduction in individual complaints might not be a sign of success, while citizen satisfaction can also be misleading as ombuds offices are primarily designed to tackle maladministration and not to ensure a contented citizenry. Disgruntled people may nonetheless have been subject to a fair process. A 2008 study in Belgium found that the effect of ombuds offices in strengthening citizen trust in government was “limited at best”; there appeared to only be a weak correlation between satisfaction on the part of those who lodged a complaint with the ombudsman and those individuals’ trust in institutions (Van Roosbroek and Van de Walle).

All this has led to assertions that “ombudsman institutions are introduced on premises that are, at best, theoretically (and not empirically) investigated” (Van Roosbroek and Van de Walle 2008: 291). Indeed, given that “effective methods to evaluate the overall impact of ombudsman remain unknown” (Stuhmcke 2014: 1), it is not surprising that there is a dearth of robust empirical evidence of their effect on integrity.

Most of the existing evaluations of ombuds offices have been conducted by the offices themselves and tend to rely heavily on output and process data, such as the number of complaints lodged (c.f. Kucsko-Stadlmayer 2008; Seneviratne 2002; Buck et al. 2011). All this leads Stuhmcke (2014: 3) to conclude that “evaluating programs like the ombudsman institution is an exceedingly difficult and uncertain endeavour”.

These considerable limitations have not stopped some observers from making very positive claims about the impact of ombuds offices in curbing corruption. Bhargava (2014), for instance, states that “ombudsman can and do fight corruption successfully when they have the enabling environment and leadership”. The International Ombudsman Institute (2022: 3) claims that ombuds offices have a “unique set of tools and competencies” that have rendered them a “useful ally” in anti-corruption efforts.

A more sober reflection indicates that the evidence remains thin and difficult to parse, especially given that many ombuds offices have assumed functions more commonly associated with anti-corruption agencies, including investigation. Today, some ombuds offices have a mandate that explicitly includes anti-corruption roles, including oversight of senior officials, assessing asset and income declarations and other such preventive tasks (Bhargava 2014). The OECD (2018b) found that 38% of the 64 ombuds offices who responded to their survey viewed themselves as contributing to anti-corruption efforts. Indeed, in 2022 the International Ombudsman Institute published a paper entitled Hybrid Corruption Ombudsman.

One of the very few studies of the effect of ombuds offices in reducing corruption was by Moreno (2016). In her empirical examination of 17 Latin American countries between 2000 and 2011, she finds some evidence that certain features of ombuds offices (Defensor del Pueblo) correlate with lower levels of corruption. Despite the fact that some observers in the region had dismissed ombuds offices as irrelevant to corruption control due to their lack of sanctioning powers, Moreno contends that ombuds offices in Latin America have been able to circumvent lengthy and compromised investigations by “appealing directly to the public and external actors” (Moreno 2016: 127). By acting as a kind of fourth estate, public appeals by ombuds offices are theorised to place pressure on government institutions to comply with accountability standards without having to resort to legal sanctions.

The key variable Moreno draws on to support her case is the “reach” of the ombuds offices; in other words, the number of regional offices each national ombudsman maintains. She found that countries with zero or very few ombuds offices were associated with higher levels of perceived public sector corruption than countries in which ombuds offices operated multiple sub-national branches to reach citizens. The regional average was 16 national and satellite offices, while in one country the national ombudsman operated 38 sub-offices with whom citizens could lodge complaints. Moreno (2016) concludes that unlike the number of complaints processed by ombuds offices, which did not appear to have a significant impact on the level of perceived corruption, “the number of satellite offices has a significant and positive effect by reducing corruption” (Moreno 2016: 141). This study therefore suggests that expanding citizen access to non-judicial methods of dispute resolution is associated, in Latin America, with lower levels of corruption.

It is important to bear in mind, however, that correlation is not causation. Plausibly, countries with a lower incidence of corruption are generally better governed, and the roll out of ombuds offices to citizens is a symptom of more democratic forms of government. Moreno (2016) herself concedes that “the strength of the office to act as an effective deterrent to corruption is endogenous to the system and actors that created this office” (Moreno 2016: 126).

Public awareness of the existence and functions of ombuds offices is vital to their work: if citizens do not know that they can report grievances to an ombuds office, that office cannot seek redress. As such, public awareness is one important – if partial – measure of the effectiveness of ombuds offices. Particularly given that a core rationale for establishing ombuds offices is to offer a more straightforward means of addressing grievances than judicial procedures, the uptake of ombuds services of poorer people who cannot afford litigation would be worth considering (Van Roosbroek and Van de Walle 2008).

Codes of conduct

According to Transparency International's definition, a code of conduct is a “statement of principles and values that establishes a set of expectations and standards for how an organisation, government body, company, affiliated group or individual will behave, including minimal levels of compliance and disciplinary actions for the organisation, its staff and volunteers” (Transparency International 2009: 8).

A frequent distinction is made between “aspirational” and “rule-based” codes of conduct (Whitton 2009). While aspirational codes establish broad ethical principles for employees, they generally do not list prohibited kinds of behaviour or set out sanctions for violations of the code (Bruce 1996). Aspirational codes of conduct are based on the assumption that unethical behaviour is largely driven by ignorance and presenting staff with information and encouragement will be sufficient appeal to their better nature to avoid integrity breaches (Meyer-Sahling and Mikkelsen 2020). Generally speaking, staff are expected, of their own volition, to aspire to the standards stipulated in aspirational codes, with support provided in seminars and workshops.

Rule-based codes are more legalistic, specifying and prohibiting inappropriate behaviours as well as providing enforceable sanctions for contraventions of the code (Bruce 1996). Whereas aspirational, peer-regulated codes are the norm in the private sector (for example, the United Nations Global Compact), public sector codes are more likely to be rule-based to enforce compliance (Transparency International 2012). Indeed, adherence to these codes is normally a condition of ongoing employment and can be made legally binding, for example through incorporation into civil service legislation (OSCE ODIHR 2012).

By now, most codes of conduct for public officials incorporate aspects of both models into a single document, often broken down into three major sections: general ethical principles, detailed provisions specifying unacceptable behaviour and a regulatory framework laying out enforcement mechanisms (Powers 2009; Bacio Terracino 2019).

Codes of conduct for public officials are very widespread and are used to cover a whole range of public servants by tailoring the codes to the specific ethical concerns and challenges the various types of public servants face in the course of their duties (Bacio Terracino 2019: 81). They are often designed to regulate the behaviour of civil servants, who often operate independently of legislators due to the need to isolate them from political influence (Rohr 1989).

Codes are believed to work in a number of ways. Firstly, they establish a benchmark to assess officials’ behaviour against the values of integrity, honesty, impartiality and objectivity (Amundsen 2009: 6; Chêne: 2013). They can also limit the pressure that supervisors and political leaders can put on public officials to act contrary to the code.

Secondly, given that not all unethical behaviour is technically illegal, codes of ethics are valuable as they can provide clarity on ambiguous points (Martini 2012). Functioning as general reference guides for officials, they offer guidance on how to deal with ethical dilemmas and outline expected standards of behaviour (Lindner 2014).

Third, they can serve as an overarching integrity management framework by formalising definitions, procedures (such as conflict of interest resolution and asset declaration) and enforcement processes.

The potential of codes of conduct has been recognised for some time, and an important step was taken in 1996 when the UN General Assembly adopted the International Code of Conduct for Public Officials and recommended that member states use it to develop their own guidelines. Since then, multilateral initiatives on codes of conduct have proliferated, notably in the United Nations Convention against Corruption (article 8) and the African Union Convention on Combating Corruption (article 7).

There is some suggestion that codes will be more effective where there is a possibility for staff to seek clarifications from an impartial ethics body or contact point who administers the code and can provide guidance (Bacio Terracino 2019; Palidauskaite no date). The existence of an independent ethics body can also help prevent the enforcement of the code from becoming a partisan political tool. The OECD (2011) notes that establishing a specific administrative structure with a mandate to oversee the implementation process is a precondition for a code of conduct's success. Other studies have likewise recommended assigning responsibility for the overall public ethics framework to one central body to oversee public officials’ codes of conduct (Reed 2008). Witton (2001) has recommended that such bodies should be empowered to audit risks to the integrity of important processes in public life such as tendering, financial management, recruitment, promotion, dismissal and discipline.

Naturally, public officials must be aware of the detail of a code’s provisions for it to be effective. Therefore, dissemination and training is considered essential to ensure that officials understand the regulations, their obligations and the standards they are expected to comply with. A 2005 study demonstrated that codes of conduct are most effective when ethical standards are clearly known as officials are then more likely to identify and denounce wrongdoing and are themselves less likely to behave in an unethical manner (Gilman 2005). Finally, Michael and Hajredini (2011) contend that compliance rates can be improved through the creation of incentives for public officials to behave ethically. Approaches include linking adherence to codes of conduct to performance evaluations and the introduction of the code of conduct during appraisal interviews for public sector jobs.

The empirical evidence on whether the existence of codes of conduct influences levels of corruption is mixed (Bacio Terracino 2019; OECD 2009: 35). A meta-review of the literature by Kaptein and Schwartz (2007: 114) found little agreement in terms of the effectiveness of ethics codes in the private sector (see Table 3).

Table 3: Findings of existing empirical studies into the effectiveness of business codes (reproduced from Kaptein and Schwartz 2007: 114)

The volume of scholarly attention devoted to the effectiveness of corporate ethics codes has remained high since Kaptein and Schwartz’s 2007 review. McKinney and Moore (2008) designed a survey questionnaire asking US business leaders to rate a vignette describing a situation that would violate the Foreign Corrupt Practices Act. They found that employees in US companies that have introduced a code of ethics were significantly less likely to view international bribery as acceptable. Halter et al. (2009) conducted a survey of 30 suppliers to a multinational corporation located in Brazil, and their results suggest that those suppliers who displayed greater awareness of the corporation’s code of ethics acted with greater transparency, leading them to conclude that codes of ethics can be a useful means of reducing corruption.

Conversely, an empirical study by Kaptein (2011) based on a sample of 2,390 adults working at US organisations employing at least 200 people found that more frequent promotion by companies of their corporate codes of conduct was associated with higher degrees of unethical behaviour. In the view of the author, however, this might be explained by the fact that in the aftermath of an integrity breach, a firm might seek to remind its employees of their responsibilities as set out in the code of conduct.

The majority of empirical studies on the effectiveness of codes have deployed surveys asking respondents in different companies about their perceptions of the effectiveness of the ethics infrastructure (e.g. McKinney et al. 2010; Schwartz 2004) or their behavioural intentions (e.g. Ruiz et al. 2015). Typically, these studies only tested for the existence of codes of conduct rather than analysing their content. This, together with a reliance on perception data, can make it difficult to understand which features of a code’s design and content are linked to enhanced effectiveness.

Kotzian et al. (2021) departed from this standard approach by using a research design that involved a between-subject experiment (factorial survey) to test the effect of a code’s tone and whether it was signed by the company’s executive board. They find that while the mere existence of a code is significant and increases the stated intention of staff to behave ethically, this positive effect is most pronounced in cases where the code bears the signature of top company executives (Kotzian et al. 2021). Conversely, a code’s tone (either positive/prescribing or negative/prohibiting) had little impact.

Finally, a meta-analysis of 100 empirical papers studying corporate codes of ethics conducted by Babri et al. (2021) found that the majority of the studies they considered had identified a positive relationship between codes and ethical intentions and behaviours. They concluded that codes “seem to be effective in terms of controlling unethical behaviour to a limited extent” (Babri et al 2021: 33). They also observed that implementing standardised codes of ethics across national boundaries and organisational hierarchies appears to undermine effectiveness.

In the public sector, Borry’s (2017) analysis of survey data from employees of a large American local government organisation suggests that codes of ethics contribute to less rule bending by staff. A longitudinal study by Beeri et al. (2013) of 108 employees in an Israeli regional council found that a code of ethics helped to positively shape staff behaviour and beliefs. Less promisingly, research by Garcia-Sanchez et al. (2011) uncovered no statistically significant correlation between levels of public sector corruption and whether the respective national administration had adopted a code of ethics.

While codes of conduct are useful components in any public sector integrity system, it has long been agreed that their mere existence is incapable of guaranteeing propriety in any organisation (Stapenhurst and Pelizzo 2004: 9). Monitoring codes’ implementation and enforcing the regulations requires significant oversight capacity.

For instance, a recent study by Meyer-Sahling and Mikkelsen (2020) drew on empirical analysis of individual level survey data of Polish civil servants to test to what extent these officials associate the use of disciplinary codes and codes of ethics with the prevalence of kickbacks in their workplace. They found that codes of ethics – when applied in isolation – have no statistically significant effect in reducing the perceived number of kickbacks. However, the authors find some indication that, when applied in conjunction with disciplinary codes that stipulate penalties for integrity breaches, the use of codes of ethics in a given government agency is associated with lower perceptions of kickbacks among civil servants working there, leading the authors to conclude that “appeals and threats appear to support each other” (Meyer-Sahling 2020: 20).

In other words, to be effective, codes of ethics need to be reinforced by disciplinary codes; the carrot and the stick are interdependent when it comes to influencing people’s decision-making processes (Meine and Dunn 2013). As observed by Bacio Terracino (2019: 82), there is by now a consensus that codes of conduct per se are not effective and must be embedded in a “wider integrity framework”. As such, a survey of the literature implies that while codes of conduct are a useful foundational document to formalise acceptable standards of behaviour, these codes should be linked to regulation, prohibition and enforcement regimes.

Ultimately, therefore, it appears that the ethics literature arrives at a similar position as rational actor models; that the most effective integrity systems seek to increase the (material) costs of non-compliance while simultaneously increasing the (social-normative) benefits of behaving ethically, such as self-esteem and the approval of peers.

In the final analysis, the weight of evidence does not seem to support the assertion by Six and Lawton (2013) and Lambsdorff (2015: 4) that punitive tools “crowd out” integrity led interventions. Rather, as put pithily by Meyer-Sahling and Mikkelsen (2020: 21), “corruption needs to be attacked from multiple sides using multiple tools at once”.

Integrity oaths

Integrity oaths have a long history, particularly where they are anchored in specific professions, such as the Hippocratic oath or the lawyer’s oath, or public functions, notably the oath of office. The prevalence of occupational oaths appeared to have declined by the late twentieth century, but there has been a revival in the last two decades, with integrity oaths having been introduced in various countries for engineers, accountants, financial advisers, pharmacists and teachers, among other professional groups (de Bruin 2016).

Particularly in the private sector, this renewed interest in integrity oaths appears to have been largely driven by the 2008 financial crises and business scandals (Jacquemet et al. 2021a). Since 2010, for example, employees in the financial services industry in the Netherlands have been required to take the Dutch banker’s oath (Tuchtrecht Banken 2020).Another prominent example is the MBA oath, where graduates from over 100 business schools around the world are invited to make a voluntary commitment to acting in accordance with the values of integrity and honesty in their future career.

The efficacy of such integrity oaths has recently been subject to academic enquiry. Jacquemet et al. (2018) tested the utility of integrity oaths as a means of enhancing honesty in organisations. In an experimental setting with 230 participants playing a game under laboratory settings, they found that test subjects who voluntarily sign an oath to tell the truth are less likely to lie despite material incentives to do so. Under the right conditions, oath-taking was able to reduce falsehoods by around 50%. In another online experiment, Jacquement et al. (2021b) found that requiring internet workers to voluntarily swear on their honour to tell the truth in economic decisions reduced cheating in a coin-flip experiment by 27%.

However, these encouraging results are nuanced by a subsequent experiment based on a dilemma scenario involving earned income and tax declaration designed by the same group of researchers (Jacquemet et al. 2020). This demonstrated that taking an integrity oath only affected the decision of people with weak preferences for lying; so-called “chronic liars” continued to behave unethically despite making the pledge. This finding mirrors the insights from Venema and van Gestel (2021: 224) in the area of nudging, that such integrity promotion interventions only seem to affect people who do not have strong pre-existing preferences or ingrained habits and attitudes.

In a final sender-receiver game experiment involving business students in France, Jacquemet et al. (2021a) determined that taking a solemn honesty oath was able to reduce “selfish lies”, where the liar profits at the expense of their interlocutor, but had no effect on reducing the number of “white lies”, where a lie benefits both parties. This suggests that the liars were able to rationalise their lie in the absence of a clear “victim” of their unethical behaviour (see Barkan et al. 2015).

De Bruin (2016: 16) has proposed several conditions necessary for an integrity oath to “foster professional, facilitate moral deliberation and enhance compliance”. According to his theory, these conditions are that the oath be made publicly in front of others, makes a general commitment to ethical principles, contains precise descriptions of intended beneficiaries (i.e. patients in the case of the Hippocratic oath), and finally emphasises the function that the oath-taker fulfils in society. In his view, these conditions “ensure oaths have greater moral weight and binding force than mere promise” (De Bruin 2016: 3). In practice, however, he concedes that the “inability of oaths to enhance compliance [with ethical rules and principles] is troublesome” and that there are open questions about their efficacy as an ethics management tool (De Bruin 2016: 16).

Integrity pacts

One prominent application of integrity oaths in the anti-corruption domain is the use of so-called integrity pacts, a social accountability tool intended to reduce kickbacks and inefficiencies in public procurement.

Since being developed by Transparency International in the 1990s, integrity pacts have been applied in hundreds of procurement processes in more than 30 countries (Pring et al. 2022). Essentially, an integrity pact is an agreement between a contracting authority and bidders, in which all parties pledge to comply with pre-agreed integrity standards and transparency throughout a public procurement chain from pre-tender to implementation and monitoring. The agreement commits parties to refrain from paying, offering, soliciting or accepting bribes, and from colluding with other bidders during the procurement process to thwart competition. Compliance with the IP is monitored by an independent third party, typically a civil society organisation.

As such, integrity pacts deploy a hybrid model that deploys both oath-taking at the beginning of the project with active oversight by civil society and (ideally) enforcement by authorities in cases of non-compliance by companies competing for a tender or implementing a procurement project (Pring et al. 2022). The integrity oath is considered to be an important “mutual recognition of shared obligations” as without joint commitment to integrity, the “process becomes essentially a unilateral pledge” (Basel Institute on Governance 2015: 24). Typically, companies undertake a commitment “not to seek or accept any benefit, which is not legally available”, or words to that effect (Basel Institute on Governance 2015: 110).

Integrity Pacts also have a didactic (instructive) function, to set precedents and good practice examples for high standards in procurement, and to raise awareness about “concepts and practices of integrity, anti-corruption and good governance” (Basel Institute on Governance 2015: 3).

Several evaluations of the application of the tool commissioned by Transparency International have found that it is perceived to be a moderately effective preventive mechanism against corruption (Pring et al. 2022; Basel Institute on Governance 2015; Blomeyer and Sanz 2015). Through semi-structured interviews, focus group discussions, and surveys with external experts and national procurement agencies, as well as document review and analysis of project monitoring data, positive attributions to the project were identified in the area of improved enforcement of procurement standards by governments and higher transparency in the procurement process. Less encouragingly, the tool seems not to have been able to generate increased engagement among citizens in the oversight of public procurement processes (Pring et al 2022).

There are a number of conditions that are seen as crucial for their implementation (Blomeyer and Sanz 2015):

• political will of the contracting authorities to reduce corruption
• transparency and professionalism throughout the contracting process
• external independent monitoring
• a participatory/multi-stakeholder involvement

Arguably, the logic here is somewhat circular, in that, to function effectively, the tool seems to rely on the enabling conditions that themselves are generally believed to reduce corruption. Moreover, the consensus among the various evaluators seems to be that the key to effectiveness is not the integrity pledge itself but rather the efficacy and expertise of the civil society monitors in detecting irregularities, as well as the pressure they can bring to bear on authorities to “prosecute and sanction corrupt behaviour” (Blomeyer and Sanz 2015: 9).

In the estimation of the most recent evaluators, integrity pacts are “only considered relevant in certain circumstances and should be used as part of a broader toolbox of options (including open data, detection and enforcement techniques)” (Pring et al. 2022: 59).

Academic honour codes

In the academic world, so-called honour codes intended to safeguard academic integrity are fairly prevalent. Despite the name “code”, these are essentially integrity pledges taken by students to uphold the expected standards of academic life (Western Sydney University no date).

A study by McCabe and Treviño (1993) of universities in the United States concluded that those institutions that implemented an honour code experienced 50% less cheating among students taking exams. A follow-up study in 2002 likewise concluded that honour codes are associated with lower rates of academic dishonesty (McCabe et al. 2002). McCabe and Treviño (1993) argue that honour codes increase students’ understanding and acceptance of expected behaviour. In addition, they suggest that honour codes also increase students’ perceptions of the level of honest and ethical behaviour among their peer group.

Subsequent work has supported the view that students from universities that have implemented honour codes tend to have a clearer understanding of what constitutes dishonest behaviour (Tatum et al. 2018). However, there is by now some consensus that honour codes themselves did not drive improved outcomes but rather student involvement in discussions with academic staff about the expectations and consequences of failing to uphold the code explains higher levels of academic integrity in institutions with a code (Tatum and Schwartz 2017; Dix et al. 2014).

Final thoughts

Ultimately, the body of evidence reviewed in this paper does not seem to support the assertion by Six and Lawton (2013) and Lambsdorff (2015: 4) that punitive tools “crowd out” integrity led interventions. Rather, as stated by Meyer-Sahling and Mikkelsen (2020: 21), “corruption needs to be attacked from multiple sides using multiple tools at once”.

The most promising results seem to emanate from interventions that raise the (material) costs of corruption while simultaneously increasing the (social-normative) benefits of behaving ethically. As such, integrity led interventions can provide a useful complement to direct anti-corruption measures but appear to unlikely to work if applied in isolation.

Annex 1: Simplified summary table