Contested Meanings of Corruption: International and Local Narratives in the Case of Paraguay

While the current international and transnational anti-corruption campaign (ITACC) has been successful in calling worldwide attention to the topic, several critics have argued that the term “corruption�? and the concepts that underlie it are ambiguous and that corruption and anti-corruption have various meanings. This paper empirically explores these supposedly divergent meanings by comparing the ITACC with the anti-corruption discourse in Paraguay. In order to explore not only the tensions but also possible coalitions between the ITACC and the Paraguayan discourse, I have conducted discourse analysis and constructionist interviews. The empirical exploration shows that differences, and thus tensions, exist between both levels with respect to the causes and effects attributed to corruption, as well as with regard to the ultimate goal of the fight against corruption. However, there also is a strong discourse coalition between the ITACC and Paraguay concerning concrete countermeasures, which indicates the dominance of the international anti�?corruption approach in the Latin American country. Very different actors with divergent understandings of corruption are able to act collectively against corruption via this discourse coalition, while still interpreting these actions according to their respective political agendas.


Introduction
Since the 1990s there has been a veritable explosion of the corruption discourse within the social sciences and practical policy fields; the international and transnational anti-corruption campaign (ITACC) 1 has been successful in calling worldwide attention to the topic. However, various scholars and political activists have criticized the ITACC and its underlying concepts and theoretical assumptions. Some critics argue that the project of fighting corruption has so 1 When speaking of the ITACC I take into account that international relations, as a product of discourse, are not limited to relations between states but are also produced and reproduced by transnational actors such as TI or the academic debate on corruption.
far not been very successful on a global scale (Bracking and Ivanov 2007: 295). 2 This lack of success, others argue, is due to anti-corruption initiative designers who work with presumptions that do not match the context where the respective anti-corruption initiative will be applied (Heeks 2007: 270). Kalin Ivanov (2007: 28) argues that the "global agenda diverges from local understandings of corruption in developing and post-communist countries," leading to misunderstandings about what exactly to fight. Ivan Krastev (2004: 35) asserts that different actors as various as transnational companies and local grassroots NGOs involved in democracy promotion engage in the anti-corruption movement under the umbrella of the ITACC.
However, he assumes that these activists have different goals in the fight against corruption.
Critiques of the ITACC are manifold and are articulated from different perspectives; however, they have two elements in common: First, that "corruption" is not an objective phenomenon with an obvious and unambiguous meaning. Rather, there are different interpretations of what the term "corruption" refers to, and therefore of what might constitute successful anti-corruption measures, according to the social, cultural, historical or regional context. Second, while pointing out the divergent understandings of corruption on the one hand, critics also assume some kind of coalition between these different actors on the other: they speak of the great variety of actors, who have different goals but who somehow unite in the fight against corruption. At the same time, however, these critics largely refrain from empirically exploring these supposed differences and possible coalitions.
The one discipline that has engaged in an empirical exploration of the various meanings of corruption is anthropology. According to one study carried out in El Alto, Bolivia, the term "corruption" is used by the broader population in a way which is contrary to the meaning that the ITACC ascribes to it: "Politicians are entitled, and expected, to accumulate wealth personally, and use it to benefit their social networks, including above all their family" (Lazar 2005: 223). The corruption charge in this context therefore signifies a general evaluation of how well or how badly money is being distributed through personal networks. That is, a politician might be called "corrupt" if he or she does not deviate from legal norms to fulfill clientelist demands. Blundo and Olivier de Sardan (2006: 113) examine various narratives legitimizing corruption in Benin, Niger and Senegal. They argue that in these countries corruption is condemned and legitimized at the same time -for example, through the idea that bending rules is being courteous. In this local context someone is called corrupt not when he or she breaks the law for personal gain, but rather when the personal gain resulting from the illegal prac-2 Despite the fight against corruption, which has been going on for more than two decades, even the World Bank, one of the main anti-corruption actors, perceives no significant positive change in the governance situation in the region. Kaufmann et al. (2009: 23)  tice is perceived to be exaggerated. They argue that in this context the corruption charge can be understood as a "benchmark of appropriateness" (ibid.: 133-134). Several other studies (Hasty 2005;Pavarala 1996;Pavarala and Malik 2010) also explore how the meaning of corruption varies between different local contexts.
While anthropological studies clearly show that corruption has different meanings in different contexts, explicit comparisons between the ITACC perspective and local perspectives are absent from the critical debate. Accordingly, existing studies do not provide empirical explorations of the above-mentioned critiques of the ITACC, and the implied differences between the ITACC and local discourses remain somewhat speculative. Moreover, the existing studies focus on the different local meanings of corruption and leave out a possible interplay between the ITACC and local discourses -and the resulting discourse coalitions. This paper seeks to address this research gap. It explores the divergent meanings of corruption as well as discourse coalitions, both of which are topics that neither the explicit critiques of the ITACC nor anthropological studies have so far convincingly addressed empirically. This paper explores the discursive tensions and discourse coalitions between the ITACC and Latin America, a region which is seen as having experienced "a genuine rise in corruption," as Kurt Weyland (1998: 110) wrote more than a decade ago. Paraguay, this paper's case study, is believed to be one of the most corrupt countries in the region and was ranked 154 out of 183 countries in Transparency International's (TI) Corruption Perception Index (CPI) in 2011 3 (TI 2011). 4 Given the high perceived level of corruption in the country, the ITACC presumes that Paraguay has a severe need to fight corruption effectively and that existing anti-corruption activities have so far not succeeded. Paraguay is a good case for studying tensions with the ITACC, because -in addition to the poor CPI ranking -corruption is also perceived from within the country to be an important and controversial topic. According to a survey conducted by a Paraguayan research center (CIRD) before President Fernando Lugo's election in 2008, a majority of respondents saw corruption as the most serious problem the new government would have to tackle (CIRD 2008: 8). In an interview with Transparency Watch (Paraguay's TI chapter) after his election, Lugo accordingly stated that by the end of his term his government would "have defeated, if not completely, then at least corruption's most grotesque and damaging manifestations" (TW 2008). This means that struggles over the meaning of corruption are potentially particularly evident in Paraguay. 3 The CPI has been criticized as having a business-biased perspective on corruption, for calling attention to the public sector while diverting attention from private-sector corruption and for severe methodological shortcomings (Galtung 2006;De Maria 2008;Andersson and Heywood 2009;Razafindrakoto and Roubaud 2010).
While I agree with these criticisms, the CPI serves here as an indicator for the ITACC, which assumes high levels of corruption in Paraguay.
4 The broader research project compares the ITACC with local discourses in Paraguay and Chile (which, in contrast to Paraguay, is perceived to be the least corrupt country in Latin America). However, this article presents the first part of the results and thus focuses on comparing the ITACC and the Paraguayan anti-corruption discourse.
The following research questions are derived from those critiques of the ITACC that point to the divergent meanings of corruption on the one hand and to the coalitions between very diverse actors in the anti-corruption movement on the other hand: How is corruption understood within the ITACC? How is this understanding structured? Which conceptions are enabled and which are excluded due to the theoretical choices made at this level? How is corruption understood at the local level in Paraguay, and which tensions emerge out of the dissimilarities between the local discourse and that of the ITACC? How do discourse coalitions between both contexts (the ITACC and Paraguay) enable collective anti-corruption action? In combining the focus on both the differences and the coalitions I hope to understand the interplay between Paraguay and the ITACC.
The multiplicity of meanings of the term "corruption" and the interplay between different (global and local) contexts can be captured by working with a notion of discourse theory (DT) that entails a relationist and a contextual perspective (Torfing 2005: 14). Following Maarten Hajer (2005: 300), I understand discourse "as an ensemble of ideas, concepts, and categories through which meaning is given to social and physical phenomena, and which is produced and reproduced through an identifiable set of practices." Thus, I presume that corruption is given meaning via discursive processes. I combine discourse analysis with constructionist expert interviews with politicians, academics and social activists in Paraguay in order to sufficiently reflect "agency" in the analysis.
The first section of the paper elaborates on existing critiques of the ITACC; it also points out the shortcomings of these critiques and their empirical foundations. The second section outlines the paper's theoretical approach, defining central concepts such as discourse, narrative and story line, which lie at the heart of any reality construction according to this paper's notion of DT. I also explain the methods applied in this analysis, such as discourse analysis and constructionist interviews. Lastly, I present the results. The discourse on corruption in Paraguay differs from that of the ITACC in two locally specific respects, the "orekuete" narrative and the "historical path dependency" narrative. However, there is also a strong convergence between the ITACC discourse and the local narratives, all of which refer to the ITACC when discussing concrete countermeasures against corruption. This convergence reveals the complex interplay between the ITACC and the local understandings of corruption in Paraguay.

Research Focus: Divergent Understandings and Discourse Coalitions
As stated above, the ITACC has been the subject of several critiques -for example, concerning definitional problems. Its definition of corruption is criticized either for including too many social practices in one concept (such as bribery, embezzlement or simply theft) (Bracking and Ivanov 2007: 298) or for being too narrow and excluding, among other things, moral evaluations of behavior in legal definitions (Schuller 1982;Philp 1997;Alemann 2005;Johnston 2005;Schefczyk 2005). The definitional debate at least implicitly entails the idea that cor-ruption has different meanings according to the -as I term it here -discursive background.
Even Pranab Bardhan (2006: 273), a prominent proponent of an economic approach to anticorruption, states, "of course, different people also mean different things by corruption." However, he continues with the widely used definition of corruption as the "use of public office for private gains," asserting that this is how most economists define the term (ibid.). Kalin Ivanov (2007: 28) assumes that the ITACC generally ignores the obvious multiplicity of meanings of corruption and argues, "the vague and emotive term 'corruption' has masked the gap between global and local discourses." The one discipline that has actually focused on and empirically explored the divergent meanings of corruption is, as mentioned above, anthropology. Shore and Haller (2005), Lazar Several other studies make this point too. According to Jennifer Hasty (2005: 272), "the pleasures of corruption" she witnessed as a journalist while working for a daily newspaper in India are interpreted by local actors as the other side of the coin of a "sphere of sociality." Engelsen  explores the ambiguity of social relations in India, which can often be labeled networking and reciprocity and at the same time corruption, depending on the point of view and social background of an actor.
While the above-mentioned anthropological studies provide initial insights into the divergent meanings of corruption, they are somewhat problematic because they suggest that "Bolivia," "Niger" or "India" in their entirety diverge from the ITACC's understanding of corruption. They present these understandings partially in an essentialist manner, largely overlooking the complex struggles over the term's meaning within local communities. That is, they largely refrain from exploring the struggles over what might be meant when corruption is referred to within Bolivia and other "local contexts." Obviously, Bolivia, or any other place, does not represent a homogenous position on this or any other topic. Vinod Pavarala (1996: 53) takes this into account and states that the concept of corruption among India's elite is not static: there is little agreement on the meaning of corruption, and the various meanings change over time. According to Pavarala and Kanchan Malik (2010: 70), actors' choice of the definition of corruption depends on practical issues and on the way they conceptualize the social world and their position in it.
I assume that rather than being undisputed (as the ITACC presumes) or being completely rejected, the meaning of corruption is subject to a complex interplay between the ITACC discourse and local discourses. Further empirical exploration is necessary to show how and in what ways these understandings differ between the different levels. Ivanov's and Krastev's assumption that meanings diverge between different actors but that these different actors simultaneously engage in the analysis of the anti-corruption movement also implies converging understandings and coalitions between diverse actors within the movement. Accordingly, this paper explores the interplay between discursive tensions and discourse coalitions.

Prescriptive-Normative Remark: A Skeptical Post-Development Perspective
Approaches that question the presumptions of mainstream anti-corruption research -such as the universality of the concept of corruption -face a certain rhetorical dilemma: they are easily accused of legitimizing corruption due to their supposed cultural relativist approach. Daniel Kaufmann 5 (1997: 115) has stated that analysts who point out the imprecision and inconsistency of the term "corruption" belong to the "fatalist camp," which according to him is not willing to actually fight corruption. From a very different viewpoint, Syed Hussein Alatas (1990: 183) calls those scholars who question, from a functionalist perspective, the presupposed unambiguously negative effects of corruption the "ideologists of corruption." The "abuse of power", which is closely related to the term "corruption," excludes the legitimation of corruption as such. However, the question arises as to whether one affiliates himself/herself with current approaches to fighting corruption or whether one critically questions them. David Kennedy (1999: 456) labels the critical position regarding the ITACC "anti-anti-corruption." The anti-anti-corruption position corresponds to a critical normative approach in the wider context of the development discourse: the post-development debate, which critically analyzes the paternalist nature of the global development project (Escobar 1995;Rahnema and Bawtree 1997;Sachs 1995;Ziai 2006: 196). This article adopts a "skeptical post-development" position (Ziai 2006: 207), which, in this research context, entails the acknowledgment of the problem of corruption and the need to overcome different forms of corruption. Such a position avoids the romanticization of particular ways of using power merely because they are labeled "traditional" or culturally embedded, as some anthropologically informed studies of corruption have implicitly done. Moreover, this position implies the need for reflection on the cultural dimensions of corruption, while also emphasizing that a culture around the phenomenon of corruption is the result of social practices that can change over time. It thereby avoids an essentialist description of either "corrupt cultures" or "the African" understanding of corruption. A post-development perspective on the current ITACC means critically questioning its current form, the way in which it produces knowledge, and its related practices. This has to happen without simply romanticizing "local" social practices on the one hand, and without rejecting the general idea of overcoming corruption on the other.

Discourse Analysis and Constructionist Interviews
In order to facilitate the research on tensions between the ITACC and Paraguay, I have chosen a particular notion of discourse theory (DT) to work with: that of Maarten Hajer, who engages in an analysis of "the role of metaphor, narratives, and story-lines in shaping and advancing strong and successful discourse coalitions" (Torfing 2005: 30).
Choosing DT as a theoretical approach implies several (anti-foundationalist) epistemological as well as (anti-essentialist) ontological claims. In DT, discourse is at the heart of all constructions of reality: everything has meaning only because meaning is ascribed to objects and phenomena (Carver 2002: 50). That is, corruption and anti-corruption cannot be studied independently of the spectator's perspective, and corruption exists only because discourse gives meaning to it. In opposition to a positivist philosophy of science, Carver (ibid.: 51) states that discourse is the linguistic creation of the world "rather than a representation of how the world is, the correctness of which could be tested." To him and other proponents of the linguistic turn, scientific evidence, too, is just one way among others of producing meaning (ibid.: 52).
This approach is suitable to exploring the simultaneous interplay of discursive tensions and discourse coalitions between the ITACC and Paraguay. It allows for the combination of an analysis of how discourse creates meaning with an analysis of how actors strategically use these constructions (Hajer 2005: 300). Thus, the aim of this paper is not just the analysis of rational argumentation but also the exposition of "argumentative rationality" (Hajer 2008a: 276). That is, I analyze not only the argumention itself, but also the wider context in which such an argumention is placed and which claims of validity are necessary to make a certain narrative sound legitimate (such as "scientifically proven facts"). Some analytical concepts from Hajer's theoretical approach are of central importance for the analysis of the anti-corruption discourse. According to Hajer, social phenomena are organized into narratives that serve an ordering function. Narratives are composed of a chain of statements. They constitute a sort of story, which usually has a beginning and an end. Most narratives presuppose certain background knowledge, since most stories are not told in full, but rather as shorter versions: story lines (Hajer 2005: 301-302;2008b: 216). Narratives in this sense exist not only through actual persons, who tell and thus produce and reproduce a story, but also through institutional routines and, for example, all sorts of publications, which also produce and reproduce specific narratives. Metaphors fulfill the function of naming structural problems and of connecting concrete problems to wider concepts (Hajer 2008a: 276). They always represent something else which is more complex. In this sense, "corruption" could be used to represent either the general moral decay of a society or an inefficient public administration.
Certain "identifiable sets of actors" use narratives strategically, and discourses are dynamic. There is constant struggle regarding a discourse's dominance: according to Hajer (2005: 303), discourse structuration means that a "discourse starts to dominate the way a given social unit (a policy domain, a firm, a society -all depending on the research question) conceptualizes the world." Subsequently, the institutionalization of a discourse might take place. This is manifested in certain institutional arrangements -for example, if a phenomenon is made visible through systems of measurement and reacted to through policy measures (Hajer 2008b: 213-219). If both criteria (structuration and institutionalization) are fulfilled, Hajer (2005: 303) speaks of the dominance of a discourse. However, with regard to the domi-nance of a discourse I diverge from Hajer's level of analysis in this article: while he speaks of the dominance of a discourse, I speak of the dominance of a narrative within a discourse, which occurs once a narrative structures the discourse and once it is institutionalized. That is, this paper looks at how the anti-corruption discourse is narrated and which narrative is dominant within the anti-corruption discourse.
Within a discourse, different actors who have divergent motivations for engaging in a discourse can unite through discourse coalitions, which enable them to attribute meanings to ambiguous and elusive social phenomena (ibid.: 302). These discourse coalitions facilitate collective action on a certain phenomenon. Actors strategically use discursive affinities to engage in and to pursue discourse coalitions in order to promote a political agenda that is necessarily connected to their respective narratives.
Empirically speaking, this paper focuses on the corruption discourse in global North-South relations within the context of the development discourse, which is represented here by the World Bank's (WB) and the International Monetary Fund's (IMF) discourse on corruption. Both organizations play a central role in the ITACC. For example, the WB itself states that it plays a "leading role in the fight against corruption" (WB 2000a: 1) and refers to research conducted by the IMF to rationalize its anti-corruption narrative (I elaborate further on the role of the WB in the ITACC in the empirical section). Central policy papers concerning the organizations' anti-corruption strategies are included in my analysis. The empirical exploration of narratives in Paraguay includes local academic literature, which is an important source because it condenses widely prevalent corruption narratives in Paraguay into compact story lines, as we will see from the empirical analysis. In examining this local context, I combine discourse analysis and constructionist interview methods, as these are useful in understanding "how interview participants actively create meaning" (Silverman 2007: 129). Although discourse analysis often tends towards a structuralist approach, I have conducted interviews with actors in key positions. Supplementing Silverman's claim, Hajer (2008b: 221) states that interviews can be used to learn about the concrete political dynamics around a given policy (e.g., anti-corruption legislation). Following Silverman's and Hajer's positions regarding interviews, in this analysis "official documents" are not considered more important than local interviews, since the decisive difference is not that between "official" and "unofficial." This latter dimension is captured in the theorem of "institutionalization," which already looks at the extent to which certain narratives are reproduced "officially" via publications. Moreover, recent debates in postdevelopment theory have emphasized that merely concentrating on the structural side of a discourse leads to the implicit construction of passive individuals who receive development discourse and are not capable of opposing such discursive structures (Ziai 2007: 227;Lie 2008: 119). Such a theoretical lack of agency is quite the opposite of what the postdevelopment approach aims for. I have conducted interviews with a variety of local actors including parliamentarians from distinct political parties, general staff from the public administration, "state modernizers" (high-ranking administrative staff and political actors who are involved in explicit statemodernization processes), and local activists involved in the fight against corruption and in democracy promotion. This constellation of interviewees, 6 which is mainly the product of a controlled snowball system, 7 has resulted in a discourse that refers to the public sector and which is reproduced by those actors who are involved in anti-corruption in a more or less immediate way. The combination of sources (academic literature, policy papers and the selected interviewees) represents an expert discourse on anti-corruption, which is important to note because it leaves out possible popular rural narratives of corruption which are not within this paper's scope of analysis.

The International and Transnational Anti-Corruption Campaign (ITACC) and the Economic Corruption Narrative
One of the ITACC's central narratives is based on rational-choice institutionalist (RCI) theoretical foundations and assumes that there are incentive-based causes of corruption. These exist wherever public officials find themselves in a situation where it seems rational for them to break formal (administrative) rules and to take bribes or to use their public office for other types of personal gain. Robert Klitgaard, who has developed one of the theoretical foundations of this narrative, accordingly states that Corruption is a crime of calculation, not of passion. True, there are both saints who resist all temptations and honest officials who resist most. But when bribes are large, the chances of being caught small and the penalties if caught meager, many officials will succumb. (Klitgaard 1998: 4) Klitgaard (1988: 75) has created a compact story line of the causes of corruption, summarizing them in a single formula: Corruption = Monopoly + Discretion -Accountability (C=M+D-A).
In the case of the state, corruption is caused by monopolies (that is, overarching "stateness," which leads to public monopolies over resources), by discretion (ergo deficient transparency), and by a lack of accountability. In such a setting it is more rewarding for a public servant to act corruptly than to comply with formal rules and duties. Klitgaard (1998: 4) 6 Interviews have been analyzed using MAXQDA, a software tool for supporting qualitative data analysis that allows for the analysis of large sets of qualitative interview data. 7 This means that I acquired interview partners mostly through local networks, then analyzed the selection and expanded it where necessary by approaching networks with different backgrounds (e.g., groups with politically opposite viewpoints) in order to capture a broader spectrum within the Paraguayan anti-corruption discourse. states that only "saints" would be able to resist such an incentive structure, and according to him this formula is valid everywhere, whether "in Ouagadougou or in Washington." The opposition of these two cities (the base of the Bretton Woods institutions and the capital of Burkina Faso) is a metaphor for the universal validity of this claim. "Saints," too, is a metaphor intended to validate Klitgaard's assumption that for (almost) every human being, acting according to one's personal benefit is an anthropological constant. However, although all actors are equally supposed to calculate the outcomes of their actions, there are "saints" (and therefore sinners). This combination allows for the claim to universal validity while at the same time creating the other: all humans act according to their individual benefit, but some can be labeled as saints and most as sinners. This combination of a moralist differentiation between "saints," "honest officials" and the rest on the one hand and the assumption of rationally calculating actors on the other hand is obviously contradictory, but it plays an important role in combining the claim to universality with an effective othering strategy. It facilitates a distinction between "us" (the ones who recognize this concept, the uncorrupt) and "them" (liars, the corrupt, the "fatalist camp" that promotes radically different perspectives on corruption).
Klitgaard's story line has experienced an extremely positive reception and is present in various central documents of the World Bank, whose anti-corruption approach is in turn very influential within the ITACC. The 1997 World Development Report (WDR), which the bank defines as "the approval and initial implementation of the Bank's anticorruption agenda" (WB 2000b: xiii), states, in accordance with Klitgaard, that "incentives for corrupt behavior arise whenever public officials have wide discretion and little accountability" (WB 1997: 103). It continues the economic narrative of corruption by referring to a market situation within the public administration: "Any policy that creates an artificial gap between demand and supply creates a profitable opportunity for opportunistic middlemen" (ibid.: 103). Although there has been a general shift in the WB's development discourse towards the post-Washington consensus (which emphasizes the importance of institutions for development, WB 2002), Klitgaard's story line remains present in this narrative. The latest, and still valid, WB policy paper on corruption "Strengthening World Bank Group Engagement on Governance and Anticorruption" (WB 2007: ii) refers to it: Within corruption-promoting incentive structures a supposed win-win situation emerges, from which both the bribed and the bribing actor benefit. However, according to the economic corruption narrative it is the political community that is impaired (Alemann 2005: 29). The outcomes related to corruption are perceived to be numerous and severe: On a daily basis, poor people around the world cannot access health clinics, schools, or other essential services because their public systems are unresponsive, or because they themselves cannot or will not pay bribes. Weak governance and corruption often mean that the resources that should fuel economic growth and create opportunities for the poor to escape poverty instead go to enrich corrupt elites. (WB 2007: 2) The fact that people cannot access health clinics, schools or other essential services because of regular fees that have to be paid is ignored as an alternative reason for social exclusion.
However, according to the economic narrative, the outcomes of corruption can be even more dramatic than those described above. They can include, for example, "financial and economic collapse, with disastrous consequences for poor people. In extreme cases, the failure of governance can result in widespread alienation, degenerating into violence, and in some instances, leading to failed states" (ibid.: 2).
The wide variety of possible negative outcomes resulting from corruption in this narrative, and their dramatic severity, is remarkable. Anybody who is in favor of "opportunities," "health" and "essential services" is compelled to align with this narrative. If not, one can be suspected of accepting "financial and economic collapse," "disastrous consequences" and even "widespread alienation," which leads to "violence." This passage of rhetorical statements again serves as a strategic element intended to underscore the unambiguity of corruption's negative effects. Given these outcomes, anyone who questions the effects of corruption, or the more basic assumptions of this conception of corruption, puts him-or herself into what Daniel Kaufmann (1997: 115) has termed the "fatalist" camp, which is not willing to actually fight corruption. This rhetorical escalation causes the discursive exclusion of any fundamental debate on, for example, the relativity of the harms done by corruption as opposed to those resulting from international trade structures, the social effects of privatization, or particular concepts of political economy. It also excludes any fundamental debate on the con- At first glance, the incentive approach in this narrative leads to the absence of an explicit corrupt subject position; this absence is represented in Klitgaard's reference to "Washington or Ouagadougou." Corruption is not caused by "people," but by incentive structures. However, the narrative does mention saints (and thus implies sinners), honest officials and liars: contrary to its claims of universal validity and certain anthropological constants, it suggests that there are some who are more corrupt than others. Also -between Washington and Ouagadougou -the World Bank's World Governance Indicators or TI's CPI leave no doubt about who is corrupt and who is not. Although the WB presents its approach to corruption as objective and neutral, within the context of global North-South relations it ultimately leads to the so-called developing countries being assigned the corrupt subject position.
The economic narrative claims validity by referring to its scientific character, implying universality and objectivity and therefore adherence to the "scientific imperative," as De Maria (2008: 780-781) rightly points out. The most recent WB anti-corruption strategy states that the outcomes outlined above have been identified through "research" (WB 2007: 3) and frequently cites IMF-based quantitative corruption research (such as the IMF volume "Governance, Corruption & Economic Performance" (Abed and Gupta 2003).
The economic narrative, not surprisingly, brings forward remedies against corruption, which "concentrate on incentive systems in tackling these problems" (Bardhan 2006: 344).
Accordingly, and fully in line with Klitgaard's compact story line, the 1997 WB-WDR states that in order to fight corruption one needs to "reduce the opportunities for officials to act corruptly, by cutting back on their discretionary authority" (WB 1997: 105). Moreover, "enhancing accountability by strengthening mechanisms of monitoring and punishment" is a central tenant of a successful fight against corruption (ibid.). In 1997 the WB still openly promoted its radical market-based approach, stating, In general, any reform that increases the competitiveness of the economy will reduce incentives for corrupt behavior. Thus policies that lower controls on foreign trade, remove entry barriers to private industry, and privatize state firms in a way that ensures competition will all support the fight. If the state has no authority to restrict exports or to license businesses, there will be no opportunities to pay bribes in those areas. If a subsidy program is eliminated, any bribes that accompanied it will disappear as well.
If price controls are lifted, market prices will reflect scarcity values, not the payment of bribes. (ibid.: 105-106) The report acknowledges that "abolishing taxes is not a sensible way to root out corruption among tax collectors" (ibid.: 106). However, the general idea of this economist, free-market and state-reductionist narrative is to eliminate those situations in which a corruption-favoring scenario could even emerge. A decade later, the WB's anti-corruption strategy no longer rejects governmental structures as vehemently, since it now speaks of "reforms that clarify the role of the state, reduce excessive regulatory burden, and promote competition" (WB 2007: ii). However, the idea that corruption is caused by immediate incentive structures and that the causes of corruption or the meaning of corruption "despite claims to the contrary, [are] not culture specific" persists (WB 1997: 8).
In the economic narrative, the final goal of anti-corruption activities is predetermined by the bank's articles of agreement, which are envisioned as "helping poor people to escape poverty and countries to achieve the Millennium Development Goals (MDGs)" (WB 2007: 2).
The bank's "main mission [is] to serve the poor" (ibid.: 3) and its overall purpose is "to help develop capable and accountable states and institutions that can reduce poverty, promote growth, and contain corruption" (ibid.). "[S]trengthening governance and fighting corruption are key to achieving this mission" (ibid.).
The economic narrative, and thus Klitgaard's understanding of corruption, is of tremendous importance within the ITACC. It is perceived to be the "most established idea of how to understand corruption" (Rothstein 2011: 230) and has achieved "wide acceptance in public policy circles and served as a foundation for empirical research and policy design to combat administrative, bureaucratic, and petty corruption" (Shah 2007: 237).
From the perspective of Hajer's discourse analysis, Shah 8 and Rothstein point to the structuration of the ITACC by the economic narrative. The translation of the economic narrative into policy design (as noted by Anwar Shah) and the production of knowledge that claims to be scientific both point to the institutionalization of this narrative within the international and transnational anti-corruption discourse. Its structuration and its institutionalization both indicate the dominance of this narrative within the ITACC.
Representatives of the ITACC will claim that there are many different concepts and conceptual developments within the international and transnational anti-corruption discourse.
However, I argue that, despite minor variations and changes within the ITACC over time, it is the economic narrative (represented by the compressed storyline C=M+D-A) that shapes the way the ITACC narrates corruption. This in turn predetermines the possible causes and effects attributed to corruption, and the countermeasures developed to fight it. This means that although the narrative might have evolved and expanded, its general foundations are still based on the idea that monopolies need to be reduced by introducing competition, and that discretion (despite its obvious inevitability) needs to be reduced wherever possible.
One of the main features of the economic narrative is the absence of an explicit corrupt subject position. This lends this narrative great adaptability, since anybody can align with it. It excludes the cultural or historical reasons behind corruption-promoting incentive structures and thus enables a variety of discourse coalitions, in this case with the Paraguayan anti-corruption discourse, between parties with otherwise quite divergent interpretations of corruption.

Dependency
Several corruption narratives are evident in Paraguay's academic literature on corruption and among the people interviewed for this paper. First, the economic narrative as such is very present in Paraguay and converges largely with the narrative of the ITACC. There are also two narratives that clearly diverge from the ITACC with respect to the causes and effects attributed to corruption and the ultimate aims of the fight against it. However, they largely refer to the concrete anti-corruption measures brought forward in the economic narrative.
They thus form a strong discourse coalition with the ITACC in this respect. I have labeled the first -divergent -local narrative the "orekuete narrative." The second divergent narrative views corruption as a historical legacy of the Stroessner regime and its domination of Paraguayan society through clientelist relations. I call this the historical path-dependency narrative (HPD).
8 Anwar Shah is a lead economist and the leader of the Governance Program at the World Bank Institute in Washington.

Economic Narrative
This narrative largely converges with that of the ITACC, and several parliamentarians, civil servants and public officials working in the area of state modernization drew upon it in the interviews conducted for this paper.
According to the local adaptation of the economic narrative, it is discretion in several ministries and in parts of the public administration that is the cause of Paraguay's corruption. One leading public official from the civil service secretariat referred to Klitgaard's story line when she stated that one of the causes of corruption was a lack of transparency: 9 "In Paraguay we didn't have information systems concerning the civil service, consolidated information systems." 10 A high-ranking official from the Paraguayan public procurement unit (Contrataciones Públicas) shared this perspective on corruption, referring to discretion and a lack of transparency as causes. Asked about the biggest challenge in his work in the public procurement unit, he stated, "Nobody knew what the state was buying, you know? At that time there was a big corrupt structure within the public institutions and among the suppliers and in the whole system of acquisitions, you know?" 11 Another civil servant, who works in the Unidad Técnica de Modernización de la Administración (Technical Unit for the Modernization of the Administration), which reports directly to the president and is responsible for promoting the modernization of the public administration, pointed to a "lack of control" as a cause of this universal phenomenon: "If something happens, it's because there is no control, because it's allowed." 12 Regarding Klitgaard's assumption about human nature, she confirmed his claim: "This is part of the human being, right? The human being will always look for ways to do certain things, right?" 13 She also asserted that corruption exists everywhere, it simply happens more in Paraguay than in other places because of this lack of control. 14 Both civil servants presented corruption as a technical problem that results from faulty incentive structures. Daniel Mendonca (2005: 74), the director of the Centro de Estudios Constitucionales (CEC) (Center for Constitutional Studies), explicitly refers to Klitgaard's formula in his study on corruption and also states that these incentive structures in Paraguay "create a system that facilitates corrupt practices." A very distinct perspective was presented by one parliamentarian from the far-right nationalist party (UNACE) who agreed with the economic narrative's inherent skepticism regarding state structures. He plainly stated, "Bureaucracy creates lots of bribes." 15 He added, "We bureaucratize too much, we have a load of civil servants, too big. We have 260,000 public servants in the Paraguayan state. And I think this makes things more difficult." 16 As is the case with the ITACC narrative, no clear corrupt subject position is presented in this local variation, making it open to strategic use by different actors and political agendas.
That is, the corrupt subject position can be adapted according to each of the two following narrative's "othering" strategies, as we will see below.
Congruent with this narrative, several administrative reforms -such as the 2003 reform of the public procurement system mentioned above -have aimed to reduce the level of discretion and to improve transparency and control mechanisms. The procurement-system reform standardized procurement processes for a larger number of private market participants by introducing information technology, which is in turn intended to enhance transparency.
Also, one current reform with respect to human resources in the public sector -the establishment of professional recruitment procedures for public staff in at least some ministriesis congruent with the economic narrative in the sense that it seeks to reduce the discretionary power of public officials when hiring new employees. This briefly outlined local economic narrative generally corresponds with that of the ITACC: corruption is not a specifically Paraguayan problem, nor is it explicitly connected to Paraguay's particular history or culture. The interviewees who put this narrative forward referred to themselves as "technical" staff objectively pursuing state modernization. However, as we will see, the following two local narratives diverge from the economic narrative (and from that of the ITACC) -for example, with respect to the causes attributed to corruption, the expected outcome of corruption and, more importantly, the ultimate aim of the fight against corruption. Moreover, they clearly imply specific political agendas. They nonetheless form a strategic discourse coalition with the economic narrative (and thus with the ITACC narrative), because they refer to the same countermeasures as its proponents.

Orekuete Narrative
The first locally specific narrative of corruption basically states that a social practice of exclusion within the Guaraní society is responsible for a corruption-promoting value system in Paraguay today (Morínigo 2005). According to this narrative, corruption is "an ethical problem." 17 Paraguay has a small indigenous population (approximately 1.5 percent of the entire population) but is generally perceived to be ethnically "mestizo." Nonetheless, almost the entire Paraguayan population is supposedly able to speak Guaraní, and in 60 percent of Paraguayan households it is the main language of communication (PNUD 2008: 126).
Against this ethno-linguistic background, José Morínigo, who is one of Paraguay's most renowned sociologists, argues that Paraguay's indigenous legacy is the cause of corruption: According to him, the Guaraní, the most prominent of the Paraguayan indigenous people, traditionally have two concepts of "us:" one "inclusive us" ("ñande" in Guaraní) and one "exclusive us" ("ore"). "Ñande" refers to a group concept that can include new members. The concept "ore," on the other hand, is an exclusive "us," which largely rejects new members and conceptualizes strong divisions between "us" and "them" (Morínigo 2005: 149). Family ties are the most important indicator of membership within the exclusive realm of the "ore," and behavior in social relationships is valued on the basis of reciprocity (ibid.: 155). The exchanging of favors between family members and -in the modern context -party members thus becomes a "moral and ethical obligation, which influences the very public image of the involved actors" (Morínigo 2009: 363). A member of parliament (MP) from the liberal party (PLRA) supports this view and describes his personal experience with such expectations of reciprocity, referring to his voters: "They practically demand us to exert influence, the John Does. (…) unfortunately, the parliamentarian who does not do this loses his people." 18 Thus, the Guaraní-based political culture is seen to clash with modern statehood and ultimately lead to corruption: "The heart of the problem lies in the value system that has been imposed on this country based on the 'ore' culture" (ibid.: 383). Morínigo calls these practices of political domination "orekuete" and claims that they are the foundation of a specifically Paraguayan system of clientelism (Morínigo 2005: 157;2009: 361). 19 If, for example, a city's mayor prefers to comply with legal norms and principles rather than getting a job for a fellow party member, he will likely be labeled "ungrateful" and a "traitor" (Morínigo 2009: 363). An MP from Patria Querida, a fairly new and morally conservative party, views this Guaraní-based value system as promoting dishonesty: "In order to be a politician one has to be a liar and dishonest. Well, this is the cultural paradigm in Paraguay; a politician has to be like this." 20 In this narrative "orekuete" is a metaphor, in Hajer's sense, for Paraguay's lack of development, due to the cultural heritage of Guaraní society, and serves as a signifier of the backwardness of indigenous value systems. The narrative concludes that because Paraguay's political process and its public realm generally function along the lines of inclusion and exclusion ("ore"), Paraguayan society is (still) not compatible with universalist concepts of modern statehood, democracy, and the rule of law (Morínigo 2005: 158). The immediate outcomes of this indigenous heritage are "nepotism" and impunity in cases of bribery. Also, "absolutely all the public institutions are overpopulated by civil servants" since it is not objective and ra-tional factors that guide the allocation of jobs in the public sector but rather traditional norms of reciprocity (Morínigo 2009: 374).
The orekuete narrative is frequently used in sociopolitical analyses of Paraguay's corruption problem. José Zanardini (2009: 631), an anthropologist, engages in this narrative; however, he does so from an empathetic perspective. The narrative is articulated in several publications -for example, in an oft-cited analysis of Paraguay's political culture by Alejandro Vial (2003: 47), who uses Morínigo's orekuete concept to explain why "the state is increasingly considered an instrument to serve private interests." Milda Rivarola, who almost became President Lugo's foreign minister, also refers to the orekuete metaphor to explain Paraguayan clientelism. Orekuete "offers advantages, privileges, and immunities to group members within a context of reciprocal obligations" (Rivarola 2007: 116). Even Roberto Céspedes (a wellrespected sociologist with the UNDP in Paraguay), who rejects the reduction of corruption to a "culturalist" explanation, mentions Morínigo's concept when explaining the cultural factors behind corruption: "The culture of a society with respect to corruption in the public sector is a key factor in determining this particular type of corruption" (Céspedes 2001: 700-701).
Interestingly, the UNDP's Human Development Report on Paraguay also refers to the "orekuete" as "cultures of inequality" that "discourage the strengthening of statehood (when they tolerate inefficiencies of the state apparatus or encourage corrupt practices, etc.)" (PNUD 2008: 74). The broad acceptance of this narrative points to the remarkable structuration of the local discourse on corruption in Paraguay, because very different actors, with different regional, socioeconomic and political backgrounds (left to far right, in the sense of progressive and conservative), use this discourse to make sense of corruption -at least with respect to its causes. However, this narrative has not been institutionalized in the sense of measurements (of, for example, the level of orekuete or reciprocal obligations) or concrete policies and programs to fight corruption.
The corrupt subject position is not evident within this narrative since it speaks of a corrupt culture and general societal value systems. However, the narrative's proponents basically state that the indigenous people and thus also the Paraguayans in general, given their ethnic background, have a tendency towards corrupt practices.
The ultimate aim of any anti-corruption activity in this narrative is "the shifting of the ethical value structure of Paraguayan society; in the sense that one is finally able to definitely uproot the practice of the 'orekuete' in the handling of public affairs" (Morínigo 2009: 383).
Thus, the aim is to change the traditional, indigenous value system to one of modern statehood. Measures to overcome corruption are described as long-term activities, because "modifying this situation in a decisive way is only possible with long-term and continuous work precisely on these values" (ibid.).
When asked what Paraguay would look like if corruption had disappeared, interviewees revealed, on the one hand, part of their present understanding of corruption. On the other hand, their answers strongly related to their respective political agendas. A parliamentarian from the liberal party (PLRA) imagined, "We would be the, I don't know whether someone says 'South America's Switzerland'." 21 Patria Querida's MP said, "I think we would be the world's prime meat-producer, the world's prime soy-producer, the world's prime producer of renewable energy and we would probably have the best diplomacy in the world." 22 He added: "I think we would be, we would be a power, I really, really believe so." Thus, the shift towards modern statehood would lead to the realization of Paraguay's actual greatness, which has been constrained by the indigenous cultural legacy of the country. However, these ultimate aims were not shared by all the actors who subscribed to this causal analysis. The idea that the indigenous heritage has led to high levels of corruption in Paraguay was distributed over different political currents, but this nationalist perspective on the final goal was only referred to by some actors from the political right.
One type of remedy, suggested by conservative proponents of this narrative, for overcoming corruption, theft, and lies is based on the individual and personal virtues one has to develop: "Our duty is a duty of abstention," stated the Patria Querida MP. "Being an honest civil servant and being an efficient civil servant is the most direct way to fight corruption." 23 Equally important to him was the principle of "not being populist." His solutions to the problem of corruption were simply the exact opposite of its causes: in his view, moral integrity is the solution to moral decay.
Interestingly, the concrete and immediate measures that these actors suggested to fight corruption form a discourse coalition with the economic narrative, and thus with the ITACC.
For example, the Patria Querida parliamentarian stated that "if we want the country to function we need to make the institutions work." 24 For Morínigo (2009: 383), too, the practice of "orekuete" must be replaced "with different values, such as transparency, efficiency and the integrity of the public administration, and (…) the idea that the 'public' is a good of all and not only of those who exert power (the rulers)." He demands "a real reform that is focused on administrative procedures and that brings real autonomy and institutionalization to the administrative machines, removing them from the control of political power" (ibid.: 382).

Historical Path Dependency (HPD): Corruption as a Legacy of the Stroessner Regime
In this local narrative of corruption neither Paraguay's specific indigenous heritage nor the related moral decay is seen as the root of the corruption-facilitating political culture. Instead, the narrative refers to the country's recent history -in particular the Stroessner dictatorship, which lasted from 1954 to 1989 -to explain corruption. Several academics have attributed the Stroessner regime's stability to the so-called "granitic unit" of the ruling political party (ANR/Colorados), the military, and the state (Lambert 1997: 6;Céspedes 2001: 699;Molinas et al. 2004: 73). According to the HPD narrative, the democratic transition since the 1990s has not been able to overcome this structural problem, which involves illegitimate connections between different powers that ought to be separate in a democracy: "Paraguay suffers from a structural situation inherited from authoritarianism and the [accompanying] permitted impunity, which operates as a mechanism for maintaining the elite" (Martini and Yore 1998: 216). Accordingly, the head of a local grassroots NGO that engages in democracy promotion stated, The state was created for this. (…) That is, this is the culture that the Colorado Party in- A high-ranking public official who works in the civil service secretariat agreed with this interpretation and added, "the sinecure, the clientelism, the favoritism, the cronyism. (...) here we have all these elements that Paraguayan public administration relied on during all this time." 26 She described the public sector's recruitment system as follows: There wasn't a system of meritocratic recruitment that was based on impartial parameters -established, transparent, available to the entire citizenship. The entrance to public employment was facilitated almost exclusively, in practically 100 percent of the cases, through political contacts, family contacts or contacts of friendship. 27 On the one hand, there is a clear discursive affinity with the economic narrative when the public official refers to discretion and a lack of transparency as a cause of corruption. On the other hand, the reference to "family" and "friends" relates to the orekuete narrative. However, both interviewees view the historical context of the Stroessner dictatorship as the reason behind the corruption-promoting incentive structures and the reciprocal value systems.
Rather than pointing to inefficiencies in the public administration that might hinder eco- resources for private ends, corruption depletes the coverage of state services, fueling inequalities" (PNUD 2008: 38, 169). Additionally, by facilitating extraordinary privileges, corruption establishes "castes of citizens that live on the margins of the law, [and a] 'marginality' from which they increase their wealth and their political and social position, which in turn allows them, at the same time, to pursue their impunity" (ibid.: 45).
The corrupt subject position in this narrative is held mainly the Colorado Party, which in 2007 still had more than 1.6 million members, half of the constituency. This construction of the corrupt subject position is the most explicit of all the narratives discussed. It allows for the most clear-cut distinction between who is corrupt and who is not, since the corrupt value system is represented by the concrete organization of the Colorado Party and its members.
This narrative is brought forward and used strategically by members of the Lugo coalition, and its proponents clearly differentiate themselves from the political actors of the Stroessner era.
When asked what Paraguay would be like if there was no corruption, the interviewees who subscribed to this narrative revealed quite different final aims than those who subscribed to the orekuete narrative. One parliamentarian who belongs to the Lugo coalition stated, Hopefully some day we will really become a country without corruption, because then we would not have children in the street, and we'd have free education, high school, university. We'd have an agrarian reform for those from the countryside, with technical assistance. We'd have universal health coverage, free assistance. We would have a paradise. 29 This vision of a Paraguay without corruption is one in which the state fulfills a broad range of public duties and needs. The statement clearly points to a specific political agenda, which is promoted through this narrative and aims to achieve social inclusion and democratization, and to overcome the oligarchic structures inherited from the Stroessner dictatorship. A civil servant working in the Ministry for Health imagined a slightly less ambitious goal in the fight against corruption: "A Paraguay without corruption? That everybody has their basic needs fulfilled. It's possible, we are few. (…) That is, here people eat garbage, you've seen that, right? Basic things. That children play and that they don't work." 30 A functioning state, the product of a successful fight against corruption, would address the basic needs of the Paraguayan population.
Remarkably, most of the concrete countermeasures presented in this narrative are congruent with those of the economic narrative. Accordingly, the public official in the civil service secretariat claimed that one would have to "combat discretionality head on," that the recruitment of civil servants would have to be organized through public competition, and that the high-ranking staff in the civil service would have to receive training in order to be "very well prepared for all they have to do." 31 In short, while this narrative clearly attributes corruption to the dictatorship and the incompleteness of the democratic transition process, and while the corrupt subject position is fairly clearly defined, the actors who adhere to it refer to countermeasures that are congruent with the economic narrative. The actors bringing forward this narrative thus -like those who endorse the orekuete narrative -form a partial discourse coalition with the ITACC. In other words, the concrete countermeasures proposed by the economic narrative (and the ITACC) serve their respective political agendas.

Discursive Tensions and Discourse Coalitions between Paraguay and the ITACC
As critics of the ITACC have stated, there are different ways of making sense of corruption, and the "agendas" behind the different definitions, as Kalin Ivanos states, are divergent.
However, several discourse coalitions, through which actors from very distinct backgrounds can relate to the economic narrative on corruption, exist between the local level in Paraguay and the ITACC. As discussed above, this economic narrative does not define the corrupt subject position clearly, nor does it openly reveal its inherent political agenda. The technocratic approach to corruption that is part of the economic narrative and the respective countermeasures allows for the adaption of this narrative to fit quite divergent political agendas. For example, actors in Paraguay who defined themselves as "technical staff" working in the area of public-administration modernization referred to this narrative and made the same causal analysis it does, stating that corruption in Paraguay does not have any specifically Paraguayan reasons, but rather that there simply exist incentive structures that promote corruption.
In contrast to the economic narrative's universalist explanation of corruption, the orekuete narrative states that the cause of corruption is the specifically Paraguayan cultural configuration of the strong influence of collective but exclusive values of reciprocity based on the indigenous culture of the Guaraní. This cultural influence is interpreted either as leading to a value system that runs contrary to the value system required by modern statehood or as a state of moral decay in which the indigenous value system causes lying, cheating, and corruption. Corruption in this narrative is thus the expression of the country's perceived backwardness. The ultimate aim in fighting corruption is to achieve modernity and -in a nationalist variant of this narrative -to fully unleash Paraguay's greatness as a nation.
The second locally specific corruption narrative also points to the specifically Paraguayan context but clearly diverges from the orekuete narrative. Here, it is the sociopolitical context of the Stroessner dictatorship, where illegitimate interconnections between the state, the ruling political party, and the military led to clientelism, cronyism, and corruption, that is the root of 31 Interview with public official from the Civil Service Secretariat, 10 May 2010, Asunción.
corruption. Corruption is understood as an instrument of political domination, and the corrupt subject position is clearly identifiable: it is limited to members of the former Stroessner system.
The ultimate aims of the respective locally specific narratives differ accordingly: Both seek a cultural shift in the public sector, but while the orekuete narrative promotes a long-term shift towards a modern value system and individual virtues, which will lead to the transformation of Paraguay to an internationally respected power, the HPD narrative seeks to overcome the authoritarian past and several related dimensions of social and political exclusion. In the latter narrative the fight against corruption represents the ongoing political struggle for democracy and the struggle against the former political, economic, and military elite.
With respect to structuration and the institutionalization of measures against corruption, the economic narrative, which is the most institutionalized, is clearly the most dominant in the Paraguayan discourse. This dominance is also demonstrated by the absence of alternative means of fighting corruption: although two narratives highlight the local specificity of the corruption problem and although they pursue quite different political agendas, they do not demand local solutions. The concrete countermeasures proposed are largely congruent and compatible with those put forward by the ITACC, thus indicating the existence of complex discourse coalitions between the ITACC and local political actors.
This also means that aligning with concrete RCI measures against corruption serves different political agendas at the local level in Paraguay. The countermeasures that are part of the economic narrative are intended to facilitate macroeconomic growth as a means to and an end of development. The same institutional engineering measures against corruption are also put forward by proponents of the orekuete. However, in this narrative they serve as a means to overcome the traditionalist/indigenous social patterns of political domination and as an instrument for facilitating a basic shift in the Paraguayan value system towards modernity, thereby strengthening Paraguay as a nation. The HDP narrative interprets these countermeasures as a means of gaining political ground against the Stroessner-related oligarchy, as represented by the "Colorados." The countermeasures proposed by the economic narrative, such as fostering transparency and reducing discretion, will serve to open up a formerly exclusive and authoritarian political system and public sector.
Far from being unambiguous, the meaning of the term "corruption" in Paraguay varies according to each narrative and the related political agenda. Accordingly, the results from this study support the idea of "diverging agendas," as Kalin Ivanov puts it, in the definition of and fight against corruption. However, as outlined above and in accordance with Ivan Krastev, they also reveal discourse coalitions between the different actors who are pursuing these divergent agendas.

Conclusion
This paper's point of departure was the ITACC and the recent critiques that scholars have raised regarding its universalist understanding of "corruption." Different strands of criticism either explicitly point out or at least imply that meanings of "corruption" differ between the main global anti-corruption actors (such as the WB or TI) and anti-corruption fighters in countries that are subject to anti-corruption efforts. While some anthropological studies have empirically investigated the contingency of meanings of corruption, they at times adopt an essentialist view of local communities, neglecting local struggles over meanings. Explicit comparisons between local perspectives and that of the ITACC, to explore not only differences but also coalitions, have largely been absent. This paper's comparative exploration has demonstrated that in Paraguay the anticorruption discourse represents political struggles over meanings and over the political agendas that are attached to the fight against corruption. The divergent narratives within Paraguay clearly show that there is no monolithic "Paraguayan" discourse and that there are competing interpretations of corruption based on the particular actors' social and political backgrounds. Two locally specific narratives are, while not strongly institutionalized, very present in the local academic literature and in the interviews. The orekuete narrative and the historical path-dependency narrative clearly diverge from that of the ITACC with respect to their causal attributions and their ultimate aims. However, almost all the concrete countermeasures proposed converge with those suggested by the ITACC, something which indicates a discourse coalition between very diverse actors. While within the ITACC the diversity of meanings of corruption is largely ignored, critics of the ITACC seem to overemphasize its paternalist nature; local actors do engage very actively and strategically in the corruption discourse, in accordance with their respective political agendas.