Towards a theory of integrity systems: a configurational approach

Abstract

The integrity of public officials is considered a key determinant of public trust in government and a central concept in good governance. An integrity system consists of all components, such as policies, practices, institutions and integrity guardians meant to contribute to the integrity of the organization at the heart of the integrity system. In this article we propose a theoretical model for the effectiveness of integrity systems that can be tested empirically. Six conditions are proposed as important for delivering the outcome of high integrity performance. Different configurations are expected to deliver the same outcome, because of varying developmental trajectories. Implications for further research are discussed.

Points for practitioners

Many countries have in place measures for dealing with corruption and unethical behaviour on the part of public officials. Rarely are these measures considered as part of a whole system. At the same time, there is little research evidence for the effectiveness of individual measures or the system as a whole. This article addresses these issues and is, therefore, important for policy makers who are developing anti-corruption measures.

Keywords

configurational analysis ethical framework ethics management integrity systems set theory

Introduction

The integrity of public officials is said to be a key determinant of public trust in government and a central concept in good governance (OECD, 2009). Many countries have put in place integrity systems or ethical frameworks to support and enhance the integrity of public officials. These ethical frameworks include regulatory measures such as codes of conduct, anti-corruption agencies or risk assessment strategies. Such frameworks sit alongside existing institutions such as audit bodies, Ombudsmen and Attorney Generals. All of these together constitute an integrity system. However, we currently know very little about the effectiveness of these systems in achieving high integrity, let alone whether public trust in government is increased.

The literature to date provides some conceptualizations of integrity systems (Head et al., 2008; Huberts et al., 2008a; Pope, 1996), but we argue that these are either descriptive or prescriptive rather than explanatory. Nor have any been tested empirically for their effectiveness. Empirical research suggests that there are different configurations of policies, institutions and practices that lead to the same results, be it high or low integrity performance (e.g. Doig and McIvor, 2003, 2008; Head et al., 2008; Huberts et al., 2008a; Quah, 2010). This suggests that variable-oriented conceptual models, that can be tested with statistical regression analyses or structural equation models, are inappropriate (Byrne and Ragin, 2010). As Doig and McIvor (2003) put it, there is no one best way.

The concept of a National Integrity System (NIS) is familiar to scholars researching integrity and corruption involving the public services and government. Coined first by Pope (1996) for Transparency International, the institutional pillars of that system, the ‘Greek Temple’, are well known (see Pope, 2000). Less well known is the concept of a Local Integrity System (LIS); some work has been done at local level in seven cities on four continents (see Huberts et al., 2008a) and a collection of essays addresses integrity systems at state level in Australia (Head et al., 2008). For national integrity systems, national leaders have to take the initiative to set up an integrity system and show the political commitment and support to make it work. International organizations or supranational governments (the IMF, World Bank, UN, OECD or EU) may put pressure on national political leaders to implement an integrity system, but it remains the responsibility of the latter to take the initiative and make it work. Local integrity systems may be imposed on local government by the central government, as, for example, was the case in the UK in 2000, or they can be local initiatives, as was the case in Amsterdam in 2001 (Huberts et al., 2008b) or Antwerp in 2004. Integrity systems may be primarily aimed at politicians (e.g. the UK), administrative officials (e.g. the Netherlands), both (e.g. New South Wales, Australia, and New York City) or politicians, administrative officials and business executives (e.g. Hong Kong).

A key question is what makes such a collection of agencies, laws and processes a system. An integrity system is a system in the sense of Bunge’s (2004: 188, in Kittel, 2006) definition: a system is ‘a complex object whose parts or components are held together by bonds of some kind’. Thus, an integrity system consists of all the components, such as policies, practices, institutions and integrity, guardians meant to contribute to the integrity performance of the government (national or local) at the heart of the integrity system. The basic characteristic of an integrity system perspective is that it outlines elements and conditions that are expected to be important for the integrity of governance. Integrity systems include both the organizational ethics management efforts within the government’s administration as well as the integrity guardians external to the administration, such as external financial auditors, ombudsmen and the police and justice system. An integrity guardian is an agency with oversight and control powers concerning integrity violations.

As well as mapping the contours of an integrity system, we are also interested in the performance of the system as a whole, and not just the individual components. Most integrity policy evaluations to date have focused on individual integrity policies such as codes of conducts (e.g. Kaptein and Schwartz, 2008), integrity training (e.g. Thorne LeClair and Ferrell, 2000), individual agencies (e.g. de Sousa, 2010; Quah, 2010) or internal investigations and sanctions (e.g. Van Tankeren, 2007). It is generally acknowledged though that the interaction between integrity policies and agencies is crucial for overall effectiveness (e.g. Maesschalck, 2005; Treviño et al., 2003). Hence, it would make sense that overall performance and results should become the focus of integrity system evaluation and not the individual policies, particularly since most of today’s societal problems, and hence governments’ tasks, are multifaceted and complex, requiring multiple policies.

The purpose of this article is to propose a theoretical model for the effectiveness of integrity systems that can be tested empirically. We start by defining integrity before reviewing prior conceptualization of integrity systems. We then develop our theory of integrity systems culminating in six conditions that are important for the effectiveness of integrity systems. Different configurations of these conditions may be found and we argue that there may be at least four different categories based on the different developmental trajectories followed over time. We conclude with implications for further research.

Integrity

The concept of integrity itself is as yet not clear and is still contested (Brenkert, 2004; Montefiore, 1999). Six and Huberts (2008) identified a range of different perspectives, ranging from integrity as wholeness to integrity as exemplary moral behavior or integrity as the quality of acting in accordance with laws and codes. We follow Six and Huberts’ (2008) arguments and work with the perspective of integrity as the quality of acting in accordance with generally accepted moral values and norms to further the public interest.

Huberts and colleagues (e.g. Huberts et al., 1999; Lasthuizen, 2008) found that nine types of integrity violations can be distinguished in the public sector, including bribery, conflict of interest, misuse and manipulation of information, patronage and cronyism. Not all integrity policies and practices target each type of violation, and Lasthuizen (2008) found very different antecedents for the occurrence of each type of violation (leadership styles, ethical climate, moral awareness). Also, when looking at the external integrity guardians, most guardians do not guard against all types of violations. Huberts et al. (2008c) showed that in the Amsterdam integrity system there are five external integrity guardians that can investigate bribes, but only one, the media, that can investigate patronage and cronyism; similarly, three external guardians guard against conflict of interest (auditor, accountant and media). In the Amsterdam system there is no external integrity agency; the Integrity Bureau is internal.

Prior conceptualizations of integrity systems

Over the years several conceptualizations of integrity systems have been proposed. The overall goal of Transparency International’s NIS-model (Pope, 2000) is to promote the integrity of governance in society. The model is built on, and sustained by, foundations which comprise public awareness and society’s values. If public awareness is high and values are strong, they will support the pillars. These pillars include the political will to fight corruption, an active parliament to attain and sustain good governance and to fight corruption, an auditor-general as watchdog over financial integrity, an attorney general as ‘guardian of the public interest’, a public service system designed to protect the public decision-making process, the judiciary and the protection of the rule of law, an ombudsman, independent anti-corruption agencies, adequate procedures for public procurement, accounting and financial management, a private sector operating within the laws and public awareness, media, civil society and international organizations supportive of ethics and integrity. The approach focuses, primarily, on institutions (Doig and McIvor, 2003).

The OECD’s (2000) focus on ethics infrastructure identifies similar actors at the national level (such as legislature, executive, judiciary, auditor-general, ombudsman, watchdog agencies and civil society), but does not utilize the temple metaphor of Pope. The OECD’s framework stresses the importance of the commitment of leaders, accountability and control, guidance to officials (e.g. through a code of conduct) and management (such as the need for a coordinating body).

The Australian NISA project (Head et al., 2008) provided two insights. First, they proposed the bird’s nest metaphor rather than the temple metaphor, since integrity systems ‘tend to rely on a diversity of individually quite weak elements for their ongoing success rather than a few strong pillars carrying a heavy load … . [T]he twigs that really make up the integrity system bird’s nest are not the individual institutions, but the cumulative interactions between the institutions’ (NISA, 2005: 8). TI’s temple metaphor suggests that the whole temple is as strong as the weakest pillar, while the Australian NISA’s bird’s nest metaphor suggests that ‘a multitude of often weak institutions and relationships can combine to more effectively protect and promote the fragile goal of public integrity’ (Sampford et al., 2005: 96). The second contribution from NISA is that they focused on three themes in their assessment: consequences, in terms of success; capacity, in terms of resources and competencies; and coherence, in terms of the system as a whole. On all three themes their conclusion is that improvement is possible and needed.

Huberts et al. (2008a) performed the first systematic comparative study of local integrity systems (LIS) for cities. They studied seven cities across four continents and began their analysis with an adaptation of Pope’s temple model, but ended with an assessment framework of nine elements. They make several conceptual contributions. First, they translate the NIS framework to local authority level. Second, their assessment framework pays more attention to system scope than previous models. The framework’s second element maps the regulatory context, in particular at the level of central government. The third element focuses on which populations are targeted by the LIS. In the UK the LIS is primarily aimed at politicians while in the Netherlands the LIS is primarily aimed at civil servants. In Amsterdam, the difference between the LIS for politicians (almost non-existent) and the LIS for civil servants (increasingly effective) is becoming an issue.

All the conceptualizations discussed above are static and institutionally focused, although many authors acknowledge the need for learning and continuous improvement over time. Integrity systems, however, operate in a dynamic context and the relations between the different parts of the system are complex. This is not sufficiently accommodated in the existing models. Also we do not know whether introduction of each model actually has a positive effect on the integrity of governance. None of the models are explanatory in nature. Pope’s and OECD’s models are largely prescriptive, and NISA’s and LIS’s models are largely descriptive. None are based on tested theory.

The empirical research to date on integrity systems shows that integrity systems tend to evolve over time along different trajectories, with different path dependencies, leading to different configurations that can show high integrity (e.g. Doig and McIvor, 2003, 2008; Head et al., 2008; Huberts et al., 2008a; Quah, 2010). The national integrity systems of, for example, the UK, the Netherlands and Hong Kong show distinct differences while the scores of the countries in international corruption databases are all high (e.g. Transparency International’s Perceived Corruption Index in 2011 for these countries is 7.8, 8.9 and 8.4, respectively). A configurational approach to integrity systems acknowledges that there may be several configurations that lead to high integrity, rather than just one best way. The configurational approach is also very promising when studying integrity systems, since it can accommodate the complex causality and nonlinear relationships that Sampford et al. (2005) found for national integrity systems in Australia. Traditional regression models are therefore inappropriate since they cannot accommodate this equifinality. Configurational analysis is also better suited to deal with the multicausality that existing conceptualizations stress than traditional regression analysis. These characteristics of integrity systems make it highly appropriate to use set theory and configurational analysis for building and testing a theory for the effectiveness of integrity systems (Fiss, 2007; Ragin, 2008; Rihoux and Ragin, 2008; Vis, 2008).

Towards a theory of integrity systems

In this section we develop a theory for the conditions that may determine the effectiveness of an integrity system. Integrity systems are considered to be effective when integrity risks are properly contained, in other words, when officials act with integrity, avoiding integrity violations. Within an integrity system, operational line agencies have primary responsibility for managing the integrity of their officials. Several actors may be involved as guardians of integrity in an integrity system: core integrity agency (also called anti-corruption agency), auditor, ombudsman, internal control, police and justice, media and civic watchdogs.

We propose that the relevant conditions to include in a conceptual model for effective integrity systems are as follows:

Conditions regarding corruption reform

Absence of corrupt institutional logic in the wider society, i.e. absence of societal values that actively support corruption.

Trigger for corruption reform, such as an ethical crisis or external pressure that leads to the formulation and implementation of new policies. This trigger may have occurred sometime in the past.

Conditions regarding external guardians

Independent oversight agencies that collectively have the power and resources (mandate, capability and capacity) to investigate, adjudicate and sanction all integrity violations that are covered by the integrity system.

Independent and free media and civic action groups that have access to public information and can publish integrity violations.

Conditions internal to government agencies

The ethical policies and practices within government agencies strengthen the internalization of values that support integrity. Each integrity risk considered important is sufficiently contained. This can be observed in:

Strong values-based policies and practices, including political will and ethical leadership.

Compliance-based policies and practices that do not weaken ethical value internalization, including internal oversight and control.

These conditions take a dynamic perspective of the systems components and do not just look at the integrity system at the moment of measurement (e.g. condition 2). They also look beyond the presence or absence of institutions to the qualities of policies and practices and operational qualities of institutions (e.g. conditions 3 and 6).

Conditions regarding corruption reform

(i) Absence of corrupt institutional logic

We see different configurations of measures and actors around the globe, and several configurations lead to high integrity performance. We hypothesize that one explanation for the presence of different configurations leading to high integrity performance is that in some countries integrity performance has always been at such a level that governments have never felt the need to start corruption reform, introducing formal rules and agencies against corruption. A strong ethical institutional logic has always been present without the need for formal rules (Misangyi et al., 2008). There was no trigger in the form of an ethical crisis, external pressure or otherwise, while in other countries low integrity performance (high corruption) at some point in time made governments embark on corruption reform, usually introducing formal rules and institutions charged with fighting corruption. The corrupt institutional logic was dominant and society’s values did not reject corruption (Pope, 2000). Formal rules and agencies for safeguarding integrity and fighting corruption were only introduced after an external trigger, a major ethical crisis or strong external pressure or regulation.

Application of this model to national integrity systems suggests four categories of countries. First, countries that have always had a dominant ethical institutional logic and therefore have had high integrity performance and have had no reasons to start major corruption reform. Consequently there will be relatively few formal agencies and rules for fighting corruption. A typical country in this category is Finland. It has consistently over time scored low on corruption in all comparative international studies, suggesting a stable dominant ethical institutional logic. At the same time, it does not have as many formal rules and agencies as many other countries (GRECO, 2004). The measures Finland has introduced focus on strengthening values internalization, i.e. the ethical logic. Thus, in countries where the ethical institutional logic has always been dominant, and who have (therefore) not had any reason to start corruption reform, there will be relatively few formal rules and agencies for fighting corruption.

A second category consists of countries that have a dominant corrupt institutional logic and have started corruption reform by putting in place formal rules and agencies for fighting corruption. They have, however, not (yet) been successful in replacing the corrupt logic with an ethical logic and strong societal values that support high integrity. A typical country in this category is Bulgaria, which scores consistently high on corruption yet on paper has many formal anti-corruption rules and agencies.

A third category consists of countries that have been successful in replacing the corrupt institutional logic with an ethical one, i.e. have undergone successful corruption reform. In the past, the corrupt institutional logic was dominant, yet at present the ethical one is dominant. What can explain their success, especially in comparison with the second category? What is the appropriate combination of policies, practices and actors to achieve high integrity risk containment and as a consequence high integrity performance? A typical country here is Hong Kong.

A possible fourth category consists of governments that overall have a dominant ethical logic within the wider society, but gaps in the risk containment system that lead to occasional ethical crises. These crises in turn trigger the implementation of new measures. Examples in this category may be the United Kingdom, the Netherlands, and also cities such as Amsterdam, New York and Antwerp. This category sits conceptually between the first and the third. The ethical institutional logic is fairly stable, supported by societal values of high integrity, but not as strong and dominant as in category one. So officials may be tempted to commit integrity violations when the risks are not properly contained. When a major scandal breaks out showing ethical misconduct by (a large numbers of) officials, there will be public demand for stricter rules, because people want better risk containment. Examples are the UK expenses scandal in 2009 but also the cash-for-questions scandal in the mid-1990s. Given the strong ethical logic that is present in society, successful implementation of these new measures will most likely be easier to achieve than in the situation of category three, where at the start of the corruption reform process the corrupt logic was strong. Similar to category three, the question is what is the appropriate combination of policies and practices to achieve high integrity risk containment and integrity performance, although it is likely that, given the different starting position, the appropriate combination may be different?

Thus the first condition is

1. Absence of corrupt institutional logic in the wider society, i.e. absence of societal values that actively support corruption.

(ii) Reform conditions

As noted earlier, new policies and practices are usually only introduced after changes in the environment. New integrity or anti-corruption measures are often introduced as part of corruption reform after an external trigger such as a major ethical crisis (e.g. Huberts et al., 2008a; Maesschalck, 2002) or external pressure, e.g. from the World Bank or the European Union (EU). A major ethical crisis is often triggered by media publishing serious ethical misconduct by (large numbers of) officials. An example is the 2009 UK expenses scandal of Members of Parliament. For this to be possible the media need to be free and independent (Pope, 2000). In some countries civic action groups exist to work with the media in investigating and publicizing such integrity violations, lobbying for corruption reform.

In Huberts et al.’s (2008a) empirical study of the development of local integrity systems it took a major crisis in each of the seven cities to get the issue of fighting corruption and safeguarding integrity onto the public and political agenda. Only then were major reforms implemented and core integrity agencies founded. In particular, if integrity performance is low, the conditions need to be present for independent and free media and civic action groups that have access to public information and can publish integrity violations.

Examples of external pressure are the EU’s pressure on new member states, such as Rumania and Bulgaria, to push through corruption reform putting in place particular actors, policies and practices of an integrity system designed by EU administrators (EU, 2008). An example at the local level is the UK where in 2001 central government imposed measures on local authorities to set up Standard Committees, appoint monitoring officers and adopt the Code of Conduct (Lawton, 2008; Lawton and Macaulay, 2004).

Thus the second condition is:

2. Trigger for corruption reform, such as an ethical crisis or external pressure that leads to the formulation and implementation of new policies. This trigger may have occurred sometime in the past.

Thus, what is the appropriate combination or configuration of integrity policies, practices and actors for each of the different categories? How can we begin to conceptualize these different components of integrity systems? What are the underlying mechanisms that determine whether a particular combination may be effective? We propose that integrity systems have two levels: external and internal. The external level consists of a network of actors with oversight mandates towards the hierarchy and these actors often have a form of horizontal accountability relationship to one another (Sampford et al., 2005).Within the internal level different actors stand in a traditional bureaucratic or politico-administrative relationship towards one another: frontline officials, operational leaders, corporate executives, corporate staff departments, and elected officials.

Conditions regarding external guardians

(iii and iv) Independent oversight agencies and media

External guardians take various forms including an ombudsman, external auditor, or a dedicated integrity (or anti-corruption) agency in various combinations. Different models of anti-corruption agencies have been identified (see Quah, 2010) with varying degrees of roles and responsibilities. Their success will depend upon a number of critical factors (Doig et al., 2007; Law, 2008). The police, media and civic organizations will also act as external guardians. Almost by definition, actors and policies operating externally from the organization at the core of the integrity system are compliance based since they are externally imposed on officials, and their effect will therefore work via external motivation and control of behaviour.

When too much emphasis is placed on external integrity guardians, the system runs the risk of reducing value internalization. The ideal is that the checks and balances internal to the operational line organisations are such that the external watchdog agencies do not need to take corrective action. Yet they are still needed to keep the targeted officials ‘on the ball’.

Thus the third and fourth conditions are

3. Independent oversight agencies that collectively have the power and resources (mandate, capability and capacity) to investigate, adjudicate and sanction all integrity violations that are covered by the integrity system.

4. Independent and free media and civic action groups that have access to public information and can publish integrity violations.

Internal conditions

(v and vi) Values and compliance-based approaches

Many integrity systems researchers appear to treat the internal level as a black box and focus on the network of external integrity guardians (e.g. Head et al., 2008; Pope, 2000). The problem with this approach is that it assumes that the key to effective integrity systems lies in the capacity and coherence of these external guardians (Head et al., 2008), while disregarding the key role of the internal-level integrity policies and practices. Other researchers more explicitly acknowledge this (Huberts and Six, 2012; OECD, 2000). What appears key at the internal level is the commitment of operational leaders to integrity management (Lasthuizen, 2008; Treviño et al., 2003), sound risk analyses and subsequent procedures to mitigate the key risks and sanction perpetrators (Lange, 2008; Maesschalck, 2005), codes of conduct (Kaptein and Schwartz, 2008) and more generally employee awareness and knowledge of integrity risks measures. It is now commonly accepted that a combination of values-based and compliance-based ethics policies and practices needs to be present for good ethics performance (Gilman, 1999; Maesschalck, 2005; Menzel, 2005; Treviño et al., 2003).

High integrity performance is achieved when integrity risks are contained; in other words, if officials act with integrity, show ethical behaviour, and avoid corruption and other integrity violations. Thus the goal of an integrity system is to steer officials’ choices for action towards ethical behaviour and away from corrupt, unethical behaviour. Tenbrunsel and Smith-Crowe (2008) propose a model for moral decision-making processes based on a central argument that decision-makers in a decision situation with ethical dimensions will either approach the situation in a moral (ethical) frame or an amoral (business, legal) frame. They base their model on extensive empirical evidence (e.g. Tenbrunsel and Messick, 1999). Both decision frames can lead to ethical behaviour or unethical behaviour (Figure 1), but there is an important distinction between the two paths leading to ethical behaviour. If officials in a moral frame decide to act ethically, they will be intentionally ethical, while if officials in an amoral frame decide to act ethically it will be unintentionally so. Empirical research shows that the probability of ethical behaviour is higher in a moral frame than in an amoral frame, ceteris paribus (p₁ > p₂ in Figure 1) (Tenbrunsel and Messick, 1999). When in an amoral frame (based on cost/benefit calculations or legal rules), officials will behave ethically only if the rules and sanctions make it the ‘right’ decision to take (Tenbrunsel and Messick, 1999). This is in line with March’s decision logic of appropriateness (1994) and Lindenberg’s distinction between self-interest master frames (hedonic and gain) versus his normative frame for decision-making (e.g. 2003).

Figure 1.

Moral decision making process (adapted from Tenbrunsel and Smith-Crowe, 2008)

Applying institutional theory concepts (cf. Misangyi et al., 2008), a moral frame is more likely if an ethical institutional logic is dominant in the organization and the wider society, while an amoral frame is more likely when a corrupt institutional logic is dominant. Lindenberg (1998, 2003) also shows that individuals are sensitive to the frames of those around them: there is ‘frame resonance’. Individuals are more likely to act from a moral or normative frame if people around them act from such a frame rather than a self-interested frame. In other words, if the values in the wider society or the organization support ethical behaviour, then a moral frame in officials is more likely, while societal values supporting corrupt behaviour make an amoral frame and thus corrupt behaviour more likely (cf. Pope, 2000).

Returning to Figure 1, the main route to high integrity is strengthening the moral frame and weakening the amoral frame (maximizing z in Figure 1). Strong values-based ethics policies and practices strengthen the moral frame (Maesschalck, 2005). Strong political will (Doig and McIvor, 2003), strong ethical leadership (Lasthuizen, 2008), ethical codes (Kaptein and Schwartz, 2008) and ethical climates stimulating dialogue about ethical dilemmas are all policies and practices stimulating the moral frame in officials. Examples of compliance-based policies and practices are rules and procedures to reduce integrity risks by taking away temptations for unethical behaviour; sanctions for unethical behaviour (and rewards for ethical behaviour) (Lange, 2008; Maesschalck, 2005). However, some compliance-based policies can crowd out the effects of values-based policies, so not every combination of policies is effective. However, there is as yet no theory to explain when a combination is effective and why. We aim to lay the foundations in this study.

Reformulating the challenge in our conceptual terms, compliance-based ethics policies and practices need to be designed and implemented in such a manner that they do not weaken value internalization, as this would weaken the moral frame (Tenbrunsel and Smith-Crowe, 2008). We use Self-Determination Theory (SDT) to provide a theoretical framework for explaining when a combination of values-based and compliance-based ethics policies and practices is likely to be effective (for an overview of SDT see Deci and Ryan, 2000; Ryan and Deci, 2000). SDT builds on earlier theories of intrinsic versus extrinsic motivation. Self-determined activities are ‘activities that people do naturally and spontaneously when they feel free to follow their inner interests’ (Deci and Ryan, 2000: 234); in other words, when they feel free to follow their own values and norms. When performing self-determined activities, people have an internal perceived locus of causality. When, however, external rewards are introduced for these activities, ‘people tend to feel controlled by the rewards, prompting a shift in the perceived locus of causality for the behavior from internal to external’ (Deci and Ryan, 2000: 234).

SDT posits that the more people internalize and integrate socially sanctioned values and norms, the more self-determined their actions and the more likely it is that their locus of causality is internal.

SDT is based on supporting the three basic psychological needs that people have: autonomy, competence and relatedness. According to SDT, it is part of the way human organisms are designed and work that they ‘want to engage [in] interesting activities, to exercise capacities, to pursue connectedness in social groups, and to integrate intrapsychic and interpersonal experiences into a relative unity’ (Deci and Ryan, 2000; 229).

The more control measures and regulations support people’s basic needs for competence, autonomy and relatedness, the more likely it is that people internalize and integrate the socially sanctioned values and norms, in other words, act from a moral frame. To act ethically, avoiding temptations to commit integrity violations, is not based on intrinsic motivation in the strict sense of the term: intrinsically motivated activities are defined ‘as those that individuals find interesting and would do in the absence of operationally separable consequences’ (Deci and Ryan, 2000: 233). Ethical behaviour, we argue, is at best fully integrated and internalized extrinsic motivation. Fully internalized extrinsic motivation ‘remains extrinsic motivation because, even though fully volitional, it is instrumental’ (Deci and Ryan, 2000: 237).

Thus, ethical behaviour and high integrity are more likely to occur with ‘high quality’ when the conditions are present for officials’ behaviour to be self-determined, which can only happen if officials are able to satisfy their basic psychological needs of autonomy, competence and relatedness. This in turn implies that all ethics measures support these needs. Values-based ethics policies and practices directly address value internalization and integration and will therefore always support the moral frame.

Most compliance-based policies are usually instigated at the corporate level, but are implemented at the operational level by operational leaders. Often, from the perspective of the street-level officials, corporate-level policies are perceived as externally imposed which implies high risk of weakening value internalization and the moral frame. Internal integrity guardians, such as internal audit and control and a dedicated integrity agency with investigative powers, are part of compliance-based measures at the corporate level.

Finally, the fifth and sixth conditions are

5. Strong values-based policies and practices, including political will and ethical leadership.

6. Compliance-based policies and practices that do not weaken ethical value internalization, including internal oversight and control.

Testing the model

What is needed now is empirical research that develops and tests the proposed theory. This research should be case-based rather than variable-based to accommodate the various configurations (Byrne and Ragin, 2010). Within the overall set-theoretical design, two methods may primarily be employed: in-depth case study analysis (typically small-N studies) and fuzzy-sets analysis (medium to large-N studies). In-depth case studies provide us with a rich understanding of the development of integrity systems over time and, where relevant, corruption reform. It will help us understand which conditions are relevant for inclusion in fuzzy-set analyses. At the same time, qualitative comparative analysis, using either crisp or fuzzy sets, will identify cases that warrant more in-depth investigation (Fiss, 2007; Ragin, 2000, 2008; Rihoux and Ragin, 2008). Hence the two methods are best applied interactively and concurrently, not sequentially.

As we find out more about what combinations of conditions, i.e. what configurations, in an integrity system lead to high integrity performance of national governments, researchers should look for the underlying social mechanisms to explain why these configurations lead to the desired outcome. This is an iterative and creative process of deepening our understanding of the individual cases and searching for explanations grounded in existing theories in other fields. Thus a theory of integrity systems is built.

Fuzzy-set analysis

Fuzzy-set analysis is a form of the set-theoretical approach. In set theory a case is analysed for its membership of different sets. For example, a case, a country, can be a member of the set of EU member states, or not. Or, a case can be considered a member of the set of democratic nations, or not. In the first example it is clear for each case whether it is a member of the set or not (crisp sets). In the second example this is less clear. Fuzzy sets allow for more variation in set membership. Fuzzy-set theory is particularly appropriate in fields where there is causal complexity. There is more than one combination of conditions that lead to a specific outcome and a certain condition can lead to different outcomes, depending on the social context (Vis, 2008). As argued above, we expect that integrity systems have causal complexity.

A key characteristic of fuzzy-set analysis is that researchers need to be very explicit and transparent about the selection of cases in relation to the research question, the theory and what they already know about potential cases. A second characteristic is the need for researchers to be very explicit and transparent about the coding or measurement of the conditions and the outcome (Ragin, 2000). Calibration of the coding of the conditions and the outcome is crucially important and needs to be based on detailed substantive knowledge of the context and the cases (Fiss, 2007). The coding of each condition and the outcome is based on the assessment of multiple indicators and qualitative insights from the cases.

The type of data required for proper measurement of the six conditions for effective integrity systems is not available in existing databases. For national integrity systems the Quality of Government dataset (Teorell et al., 2010) collates data from, among others, the Global Integrity Report (GIR), World Bank Governance Indicators (WBGI) and World Values Survey (WVS). None measure the conditions at the required level of detail, focusing on ethical policies and practices, distinguishing between compliance-based and values-based policies. For the outcome, integrity risk containment, three databases may provide useful preliminary proxies with indicators for perceptions of level of corruption (Transparency International’s Corruption Perception Index, 180 countries, from 1995; World Bank’s Control of Corruption indicator, 212 countries, from 1996; and International Country Risk Guide’s corruption indicator, 140 countries, from 1984).

The software package that may be used to perform the fuzzy-set analysis is fs/QCA. The first step of the analysis consists of constructing ‘a truth table that lists all possible configurations [of conditions] as well as whether these configurations lead to the outcome in question’ (Fiss, 2007: 1184). In the next step fuzzy-set logic is applied ‘to determine commonalities among the configurations’ and to generate a logical conclusion (Fiss, 2007: 1184). Fuzzy sets are mathematically proven (Ragin, 2000).

Hypothetical illustration

To illustrate how set analysis could work in analysing the effectiveness of NIS, we give a hypothetical illustration. For simplicity’s sake, in this hypothetical application we use crisp sets, where the presence of components is defined using binary values (yes/no). For each relevant condition of the NIS system the effective presence is coded as ‘yes’ (= present) or ‘no’ (= not present). We limit this hypothetical illustration to four conditions:

Presence of effective values-based policies and practices

Presence of independent oversight agencies (watchdogs)

Presence of effective independent and free media

Absence of corrupt institutional logic in society.

The outcome is coded for the presence of high integrity performance (= ‘yes’) or not (= ‘no’). See Table 1 for the resulting ‘truth table’ with the coding (Fiss, 2007). With four conditions and for each condition two possible values (present or absent), there are 2⁴= 16 total theoretically possible configurations. For four of these configurations no cases were found (B, N, O and P). For four configurations the desired outcome was found (A, H, J and M). In each of these four configurations condition IV, no corrupt logic in wider society, is present. This makes it a necessary condition, but it is not a sufficient condition, because in configurations D, E and Q it is also present without logically implying the desired outcome. In configuration A all four conditions are present, but in the other three effective configurations only two of the four possible conditions are present.

Table 1.

‘Truth table’

Possible configurations	Component I Values-based policies	Component II Independent watchdogs	Component III Independent media	Component IV No corrupt logic	Outcome High integrity
A	Yes	Yes	Yes	Yes	Yes
B	Yes	Yes	Yes	No	?
C	Yes	Yes	No	No	?
D	Yes	Yes	No	Yes	No
E	Yes	No	Yes	Yes	No
F	Yes	No	Yes	No	No
G	Yes	No	No	No	No
H	Yes	No	No	Yes	Yes
J	No	Yes	Yes	Yes	Yes
K	No	Yes	Yes	No	No
L	No	Yes	No	No	No
M	No	Yes	No	Yes	Yes
N	No	No	Yes	Yes	?
O	No	No	Yes	No	?
P	No	No	No	No	?
Q	No	No	No	Yes	No

The full statement reads:

(I · II · III · IV) + (I · IV) + (II · IV) + (III · IV) → high integrity performance

(Note that: ċ = AND; + = OR; → = logically implies; I = condition I, etc.)

Using Boolean logic to simplify the statement yields:

(I · IV) + (II · IV) + (III · IV) → high integrity performance

In fuzzy-set analysis, these set-theoretic relationships are evaluated by calculating the coverage of each consistent subset, an analysis ‘somewhat analogous to the partitioning of explained variance in multiple regression analysis’ (Ragin, 2006: 291). By calculating the consistency, ‘the degree to which a subset relation has been approximated’ is assessed (Ragin, 2006: 291). The software that is used for the analysis can calculate these measures as well as assess the effect of limited diversity, that is, not having cases for all theoretically possible configurations.

In the model proposed, there are six conditions and therefore there are in principle 64 (2⁶) possible configurations, but some are theoretically unlikely, e.g. those in which there is both a strong presence of values-based ethical policies and practices in government agencies, indicating strong internalization of ethical values by public officials, and a strong presence of the corrupt institutional logic in wider society (this takes away 16 possible combinations). Other configurations are empirically unlikely in today’s globalized world, e.g. those in which a strong corrupt institutional logic is present in wider society and there is no trigger for reform, in particular no external pressure from international organizations such as the United Nations, World Bank or European Union (this takes away eight possible combinations).

Conclusions and implications for further research

This study developed theory for assessing the effectiveness of integrity systems. What are effective integrity systems and why? This is urgently needed as the models proposed to date are either descriptive or prescriptive; none are explanatory. Yet many are used by international development and aid agencies in practice. We use a configurational approach, based on set theory, since (descriptive) empirical research suggests that there is no one best way; there are different configurations of policies, practices and institutions that can lead to the same outcome of high integrity performance and high integrity risk containment. One of the explanations for this equifinality may be the different path dependencies and development trajectories of the different integrity systems. We propose four categories of countries, based on their different developmental trajectories over time. We propose six conditions that contribute to the outcome, integrity performance of the system. The levels of analysis of our model are the individual, the organization and the external environment.

What is needed now is empirical research that develops and tests the proposed theory. This research should be case-based rather than variable-based to accommodate the various configurations (Byrne and Ragin, 2010). Qualitative comparative analysis, using either crisp or fuzzy sets appears appropriate (Fiss, 2007; Ragin, 2008; Rihoux and Ragin, 2008). Ideally, the theory is tested with many cases (medium to large N), but smaller-N studies may also shed valuable insights on the merits of the theory.

As countries and cities alike turn, increasingly, to integrity agencies and other ethics policies and practices to assist in the fight against corruption and integrity violations, it is timely to examine how effective integrity systems are. The innovative theory developed in this study makes substantial contributions to the field’s theory development, focusing on what effective integrity system configurations are and why. The model may point to situations where few (if any) of the conditions necessary for successful corruption reform, i.e. moving from low to high integrity performance, are present. In countries that currently have low integrity performance, outside pressure, even when financial and consulting support are provided, to introduce anti-corruption agencies, passing laws and introducing new regulations, may be futile as long as little attention is paid to the institutional logic in the wider society. This is consistent with the argument presented by Persson et al. (2013), who argue that corruption is a collective action problem rather than a principal–agent problem. Furthermore. a focus on compliance-based policies and practices without attention to the degree to which they weaken or strengthen ethical value internalization is likely to be futile is such situations.

Based on this improved theory-based understanding of integrity systems, we are confident that evidence-based practical advice may eventually be given to political and administrative leaders worldwide who want to curb corruption and improve the integrity of governments.

Footnotes

Frédérique Six is assistant professor of Governance and Organization in the department of Governance studies at VU University Amsterdam. Her research focuses on public governance puzzles in general and, in particular, the relation between trust and control. She has published in journals such as Journal of Management Studies, Public Management Review and International Journal of Human Resource Management, and has co-edited the books Local Integrity Systems (with Huberts and Anechiarico, BJu, 2008) and The Trust Process (with Nooteboom, Edward Elgar, 2003).

Professor Alan Lawton is Deputy Head of the School of Business and Economics at Monash University, Gippsland campus. His main research interests are in ethics and public service organizations and he has published extensively in this area. His latest book, with Julie Rayner and Karin Lasthuizen, is Ethics and Management in the Public Sector (Routledge, 2012).

References

Brenkert

(2004) The need for corporate integrity. In: Blenkert

(eds) Corporate Integrity and Accountability, Thousand Oaks, CA: Sage, pp. 1–10.

Byrne

Ragin

(2010) The Sage HAndbook of Case-based Methods, Los Angeles, CA: Sage.

Deci

Ryan

(2000) The ‘what’ and ‘why’ of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry 11(4): 227–268.

de Sousa

(2010) Anti-corruption agencies: Between empowerment and irrelevance. Crime, Law and Social Change 53: 5–22.

Doig

McIvor

(2003) The national integrity system: Assessing corruption and reform. Public Administration and Development 23(4): 317–332.

Doig

McIvor

(2008) Using national integrity systems to assess corruption: An international perspective. In: Head

Brown

Connors

(eds) Promoting Integrity: Evaluating and Improving Public Institutions, Farnham, Surrey: Ashgate Publishing, pp. 55–68.

Doig

Watt

Williams

(2007) Why do developing country anti-corruption commissions fail to deal with corruption? Understanding the three dilemmas of organizational development, performance expectation, and donor and government cycles. Public Administration and Development 27: 251–259.

EU (2008) On the management of EU-funds in Bulgaria. Report from the Commission to the European Parliament and the Council, Brussels, 23 July.

Fiss

(2007) A set-theoretic approach to organizational configurations. Academy of Management Review 32: 1180–1198.

10.

Gilman

(1999) Public sector ethics and government reinvention: Realigning systems to meet organizational change. Public Integrity 1: 175–192.

11.

GRECO (2004) Second evaluation round. Evaluation report on Finland, Strasbourg: Council of Europe.

12.

Head

Brown

Connors

(2008) Promoting Integrity: Evaluating and Improving Public Institutions, Farnham: Ashgate.

13.

Huberts

LWJC

Six

(2012) Local integrity systems: Towards a framework for comparative analysis and assessment. Public Integrity 14(2): 151–172.

14.

Huberts

LWJC

Pijl

Steen

(1999) Integriteit en corruptive [Integrity and corruption]. In: Fijnaut

Muller

Rosenthal

(eds) Politie. Studies over haar werking en organisatie, Alphen aan den Rijn: Samsom.

15.

Huberts

LWJC

Lasthuizen

Peeters

(2006) Measuring corruption: Exploring the iceberg. In: Sampford

Shacklock

Connors

Galtung

(eds) Measuring Corruption, Aldershot and Burlington, VT: Ashgate Publishing, pp. 265–294.

16.

Huberts

LWJC

Anechiarico

Six

(2008a) Local Integrity Systems: World Cities Fighting Corruption and Safeguarding Integrity, The Hague: BJu Publishers.

17.

Huberts

LWJC

Six

Lasthuizen

(2008b) Evaluation of Amsterdam’s local integrity system. In: Huberts

LWJC

Anechiarico

Six

(eds) Local Integrity Systems: World Cities Fighting Corruption and Safeguarding Integrity, The Hague: BJu Publishers, pp. 211–228.

18.

Huberts

LWJC

Anechiarico

Six

van der Veer

(2008c) Local integrity systems analysis and assessment framework. In: Huberts

LWJC

Anechiarico

Six

(eds) Local Integrity Systems: World Cities Fighting Corruption and Safeguarding Integrity, The Hague: BJu Publishers, pp. 271–295.

19.

Kaptein

Schwartz

(2008) The effectiveness of business codes: A critical examination of existing studies and the development of an integrated research model. Journal of Business Ethics 77(2): 111–127.

20.

Kittel

(2006) A crazy methodology? On the limits of macro-quantitative social science research. International Sociology 21(5): 647–677.

21.

Lange

(2008) A multidimensional conceptualization of organizational corruption control. Academy of Management Review 33(3): 710–729.

22.

Lasthuizen

(2008) Leading to Integrity: Empirical Research into the Effects of Leadership on Ethics and Integrity, Amsterdam: VU University Amsterdam.

23.

Law

(2008) The Hong Kong integrity system. In: Huberts

LWJC

Anechiarico

Six

(eds) Local Integrity Systems: World Cities Fighting Corruption and Safeguarding Integrity, The Hague: BJu Publishers, pp. 79–104.

24.

Lawton

(2008) Evaluation of the London integrity system. In: Huberts

LWJC

Anechiarico

Six

(eds) Local Integrity Systems: World Cities Fighting Corruption and Safeguarding Integrity, The Hague: BJu Publishers, pp. 135–150.

25.

Lawton

Macaulay

(2004) Ethics at the crossroads? The ethical framework for local government. Local Government Studies 30(4): 624–638.

26.

Lindenberg

(1998) Solidarity: Its microfoundations and macrodependence, a framing approach. In: Doreian

Fararo

(eds) The Problem of Solidarity, Theories and Models, London: Gordon and Breach, pp. 61–112.

27.

Lindenberg

(2003) Governance seen from a framing point of view: The employment relationship and relational signalling. In: Nooteboom

Six

(eds) The Trust Process in Organizations: Empirical Studies of the Determinants and the Process of Trust Development, Cheltenham: Edward Elgar, pp. 37–57.

28.

Maesschalck

(2002) When do scandals have an impact on policy making? A case study of the police reform following the Dutroux scandal in Belgium. International Public Management Journal 5(2): 169–193.

29.

Maesschalck

(2005) Approaches to ethics management in the public sector: A proposed extension of the compliance–integrity continuum. Public Integrity 7: 21–41.

30.

March

(1994) A Primer on Decision Making: How Decisions Happen, New York: The Free Press.

31.

Menzel

(2005) Research on ethics and integrity in governance: A review and assessment. Public Integrity 7: 147–168.

32.

Misangyi

Weaver

Elms

(2008) Ending corruption: The interplay among institutional logics, resources and institutional entrepreneurs. Academy of Management Review 33(3): 750–770.

33.

Montefiore

(1999) Integrity: A philosopher’s introduction. In: Montefiore

Vines

(eds) Integrity in the Public and Private Domains, London: Routledge, pp. 3–18.

34.

NISA (2005) Chaos or Coherence? Strengths, Opportunities and Challenges for Australia’s Integrity Systems, Brisbane: Key Centre for Ethics, Law, Justice and Governance, Griffith University.

35.

OECD (2000) Trust in Government: Ethics Measures in OECD Countries, Paris: OECD.

36.

OECD (2009) Integrity in Government: Towards Output and Outcome Measurement, Paris: OECD.

37.

Persson

Rothstein

Teorell

(2013) Why anticorruption reforms fail: Systemic corruption as a collective action problem. Governance: An International Journal of Policy, Administration, and Institutions 26(3): 449–471.

38.

Pope

(1996) National Integrity Systems: The TI Sourcebook, Berlin: Transparency International.

39.

Pope

(2000) National Integrity Systems: The Transparency International Source Book, Berlin: Transparency International.

40.

Quah

JST

(2010) Defying institutional failure: Learning from the experiences of anti-corruption agencies in four Asian countries. Crime, Law and Social Change 53: 23–54.

41.

Ragin

(2000) Fuzzy-set Social Science, Chicago, IL: The University of Chicago Press.

42.

Ragin

(2006) Set relations in social research: Evaluating their consistency and coverage. Political Analysis 14(3): 291–310.

43.

Ragin

(2008) Redesigning Social Inquiry: Fuzzy Sets and Beyond, Chicago, IL: The University of Chicago Press.

44.

Rihoux

Ragin

(2008) Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques, Thousand Oaks, CA: Sage.

45.

Ryan

Deci

(2000) Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist 55(1): 68–78.

46.

Sampford

Smith

Brown

(2005) From Greek temple to bird’s nest: Towards a theory of coherence and mutual accountability for national integrity systems. Australian Journal of Public Administration 64(2): 96–108.

47.

Six

Huberts

LWJC

(2008) Judging a public official’s integrity. In: Huberts

LWJC

Maesschalck

Jurkiewicz

(eds) Ethics and Integrity of Governance: Perspectives across Frontiers, Cheltenham: Edward Elgar, pp. 65–82.

48.

Tenbrunsel

Messick

(1999) Sanctioning systems, decision frames, and cooperation. Administrative Science Quarterly 44(4): 684–707.

49.

Tenbrunsel

Smith-Crowe

(2008) Ethical decision making: Where we’ve been and where we’re going. Academy of Management Annals 2(1): 546–607.

50.

Teorell J, Samanni M, Charron N, Holmberg S and Rothstein B (2010) The quality of government dataset, version 27 May 2010. University of Gothenburg: The Quality of Government Institute. Available at: http://www.qog.pol.gu.se.

51.

Thorne LeClair

Ferrell

(2000) Innovation in experiential business ethics training. Journal of Business Ethics 23(3): 313–322.

52.

Treviño

Brown

Hartman

(2003) A qualitative investigation of perceived executive leadership: Perceptions from inside and outside the executive suite. Human Relations 56: 5–37.

53.

Van Tankeren

MHM

(2007) Het preventieve integriteitsbeleid van de politie Amsterdam - Amstelland, Amsterdam: Politie Amsterdam-Amstelland.

54.

Vis

(2008) Biting the Bullet or Steering Clear? Politics of (not-)Unpopular Welfare State Reform in Advanced Capitalist Democracies, Amsterdam: VU University Amsterdam.

Possible configurations	Component I Values-based policies	Component II Independent watchdogs	Component III Independent media	Component IV No corrupt logic	Outcome High integrity
A	Yes	Yes	Yes	Yes	Yes
B	Yes	Yes	Yes	No	?
C	Yes	Yes	No	No	?
D	Yes	Yes	No	Yes	No
E	Yes	No	Yes	Yes	No
F	Yes	No	Yes	No	No
G	Yes	No	No	No	No
H	Yes	No	No	Yes	Yes
J	No	Yes	Yes	Yes	Yes
K	No	Yes	Yes	No	No
L	No	Yes	No	No	No
M	No	Yes	No	Yes	Yes
N	No	No	Yes	Yes	?
O	No	No	Yes	No	?
P	No	No	No	No	?
Q	No	No	No	Yes	No

Possible configurations	Component I Values-based policies	Component II Independent watchdogs	Component III Independent media	Component IV No corrupt logic	Outcome High integrity
A	Yes	Yes	Yes	Yes	Yes
B	Yes	Yes	Yes	No	?
C	Yes	Yes	No	No	?
D	Yes	Yes	No	Yes	No
E	Yes	No	Yes	Yes	No
F	Yes	No	Yes	No	No
G	Yes	No	No	No	No
H	Yes	No	No	Yes	Yes
J	No	Yes	Yes	Yes	Yes
K	No	Yes	Yes	No	No
L	No	Yes	No	No	No
M	No	Yes	No	Yes	Yes
N	No	No	Yes	Yes	?
O	No	No	Yes	No	?
P	No	No	No	No	?
Q	No	No	No	Yes	No

Possible configurations	Component I Values-based policies	Component II Independent watchdogs	Component III Independent media	Component IV No corrupt logic	Outcome High integrity
A	Yes	Yes	Yes	Yes	Yes
B	Yes	Yes	Yes	No	?
C	Yes	Yes	No	No	?
D	Yes	Yes	No	Yes	No
E	Yes	No	Yes	Yes	No
F	Yes	No	Yes	No	No
G	Yes	No	No	No	No
H	Yes	No	No	Yes	Yes
J	No	Yes	Yes	Yes	Yes
K	No	Yes	Yes	No	No
L	No	Yes	No	No	No
M	No	Yes	No	Yes	Yes
N	No	No	Yes	Yes	?
O	No	No	Yes	No	?
P	No	No	No	No	?
Q	No	No	No	Yes	No