Like Chameleons: Civil Servants and Corruption in Malawi 14

no abstract 


*Gerhard Anders
In the civil service you have to know how to manoeuvre.
Senior civil servant in Malawi The making of rules and social and symbolic order is a human industry matched only by the manipulation, circumvention, remaking, replacing, and unmaking of rules in which people seem almost equally engaged. Moore (2000:1) In many tales and myths of Malawi's past the chameleon figures as a morally ambiguous and cunning character (Schoffeleers and Roscoe 1985:17-38). The acclaimed poet Jack Mapanje (1981), who was imprisoned without charge or trial between 1987 and 1991, used the image of the chameleon to describe survival strategies during the days of Kamuzu Banda's authoritarian regime when persecution and detention of so-called "confusionists" and "separatists" was the order of the day. Like chameleons, adopting the colours of the background to outwit their predators, people had to veil their criticisms of the regime to avoid detention. Since the introduction of multi-party democracy in 1994 the chameleon is used as a metaphor to describe the opportunism of politicians who sell their vote to the highest bidder (Dzimbiri 1998, Englund 2002. Civil servants in Malawi resemble chameleons in many ways. In order to manoeuvre successfully in the civil service they have to negotiate the often-conflicting claims and expectations from colleagues, superiors and kin. During President for Life Kamuzu Banda's authoritarian regime between the country's independence from Britain in 1964 and the introduction of multi-party democracy in 1994 civil servants behaved like chameleons in Jack Mapanje's sense whilst since 1994 many seem to take advantage of the new liberties to use the public office for personal gains. Malawi, known until the 1990s for relatively high levels of integrity and diligence among the country's civil servants, seems to have caught up with the rest of Africa although the situation is a far cry from countries such as Angola or Nigeria where corruption is rampant. Since the introduction of multi-party democracy corruption appears to be on the rise and is hotly debated in the media and the political arena. Generally my interlocutors stressed that corruption was virtually unknown in Malawi until 1994 and expressed sentimental feelings for the "good old time" under the rule of Kamuzu Banda and the Malawi Congress Party (MCP), the only legal party till 1993. It was striking how popular discourse drew a clear line between the period between 1964 and 1994 with little corruption, on the one hand, and the time since 1994 with widespread corruption and nepotism, on the other. 15 This perceived increase in corruption since the 1990s has several reasons. 16 First of all it is now official policy to talk about it openly. Since 1994 media and people have enjoyed more leeway to talk about corruption than in the past. Newspapers, for example, mushroomed and although most of them were only shortlived a range of professional national newspapers survived. In 2002, three national newspapers were available in the urban areas and many trading centres: The Nation, the Daily Times and the Chronicle, an investigative weekly. These papers reported extensively on corruption scandals among politicians and public servants in spite of occasional harassment from the government. This is something unheard of under Banda's rule when strict control and arbitrary arrests of alleged dissidents created a culture of paralysis and fear. In those days any critical remark could lead to one's arrest or interrogation and the only newspaper, the Daily Times, primarily functioned as the mouthpiece of Kamuzu Banda and the MCP.
The changes in attitude and policy are to a large degree attributable to the keen interest the "donor community", led by World Bank and IMF, has been taking in the improvement of governance since the 1990s. The democratically elected government's alleged susceptibility to corruption has been in contrast with its publicly announced policies to curb corrupt practices. In 1995 parliament passed the Corrupt Practices Act and in 1998 the donor-funded Anti-Corruption Bureau became operational. This policy is a direct product of the promotion of "good governance" by the World Bank, the IMF and the international donor community (e.g. World Bank 2000aBank , 2000b. This is in stark contrast to the time under authoritarian rule. In the past practices nowadays perceived to be corrupt were simply part of the political system. In fact, there existed no boundaries between Kamuzu Banda's personal property, the ruling and only legal party MCP, and the state institutions. As a retired district commissioner put it: "when the district officer of the MCP wanted to borrow your tractor you couldn't refuse, could you?" This has changed considerably since 1994 and in recent years politicians and functionaries have often been accused of and have been accusing each other of corruption (Englund 2002). Corruption accusations are now routinely used to get rid of political rivals and often it is hard to tell what motivates an official investigation into the conduct of officials and politicians.
Finally, it should be noted that many civil servants cite the dire economic situation as explanation for the apparent increase of corruption. Malawi has been in the grip of economic decline since 1982, which has turned into a full-blown economic crisis since the early 1990s, when free-market reforms were being implemented by the World Bank and the IMF. Inflation has constantly eroded civil 15 Anthropological fieldwork was carried out between November 1999and November 2000, and February 2002and March 2002 This article only addresses the perceived increase in corruption since there is no reliable data on the phenomenon either before or after 1994. The only thing that can be safely said is that the talk about the phenomenon has increased since 1994. The various corruption indices only present a very incomplete picture reflecting the current obsession with corruption rather than its actual extent.
servants' salaries: in 1992 the real value of basic salaries was about 50 per cent below the levels of 1982 (World Bank 1994:37-39). Due to currency devaluations and high inflation rate this trend has intensified since 1994. 17 Poverty is widespread, endemic illnesses such as HIV/AIDS and malaria produce staggering mortality rates and civil servants are not only confronted with their own social decline but also by more demands on their resources from impoverished relatives and friends. Furthermore, there is growing discontent about the glaring income differences between them and functionaries of donor agencies and NGOs who often do the same job and have the same qualifications but who earn several times the salary of a civil servant. 18 Consequently, civil servants often refer to their meagre salaries as a justification and explanation of corrupt behaviour.
The socio-political context, which I could only sketch here, provides the background for the analysis presented in this article. It takes the ambivalence of vernacular conceptions of practices labelled as corrupt as point of departure. Behaviour violating official rules and regulations that can be justified on moral grounds is often secretly condoned by colleagues and superiors. This article argues that the ambivalence towards practices officially labelled as "corrupt" is the consequence of the existence of alternative normative orders justifying and regulating corrupt practices. Continuously switching from one normative code to the other while negotiating their course between expectations from dependants, patrons, colleagues and superiors civil servants resemble chameleons as a case study of a civil servant will exemplify. The case study is followed by an analysis of the different sets of rules either prohibiting or justifying certain forms of corruption.

Corruption in the vernacular
The introduction of multi-party democracy and anti-corruption policies since the 1990s has had a considerable impact on popular discourse. During my fieldwork people used the English word "corruption" quite frequently in everyday conversations in Zomba and Lilongwe but in a very broad lay sense, covering corruption in the legal sense of the term, i.e. bribery, and all kinds of illicit practices like fraud, theft, embezzlement, etc. In everyday conversations, newspaper reports, official documents and statements by politicians "corruption" was invariably condemned as something poisonous that had to be eradicated or stamped out. However, this general rejection of "corruption" did not imply a clear and unequivocal rejection of practices considered to be corrupt.
In many conversations people used the Chichewa word katangale. Katangale denotes any kind of illegal, dubious or shady deal or practice connected to the place of work or the office in the formal sector. Katangale also covers nepotism and 17 And indeed salaries are generally quite low: in 2002 most junior grades like security guards, gardeners and messengers earned about US $ 20 per month. Extension workers and primary school teachers earned between US $ 25 and US $ 30 per month. Even officers with higher qualifications, like a diploma or a bachelor's degree, rarely earned more than US $ 50 per month: junior professional officers as secondary school teachers, for example, had a monthly salary of $ 40. The highest grades, in the so-called Superscale, earned roughly US $ 100 per month. 18 Excluding European and US consultants hired by the World Bank who earn several hundred US $ a day.
patronage since the two practices are often connected. Katangale is a very ambivalent concept. Whilst people usually express strong disapproval when they use the English word corruption they tend to regard katangale as an excusable way of making do. Although people rarely approve of it explicitly they often talk about it in a rather ironic fashion thereby acknowledging the fact that it is an intrinsic aspect of formal employment. Katangale is not so much an individualist action but rather an aspect of "the system". According to the popular perception katangale does not constitute an act of a person's free will but something he or she is drawn into; " it is the way how things are done around here", a system that does not really leave an alternative way of doing things. This denial of agency is a common register used to justify corruption, as Blundo and Olivier de Sardan (2001) point out.
Two other words were also used to refer to corrupt practices: ziphupu and madeal. Ziphupu is a minor act within the meaning of katangale and denotes activities associated with theft or embezzlement. Madeal is also a practice that falls under the term katangale and is derived from the English word "deal". The word describes illegal and semi-illegal deals and transactions between two or more people. Katangale has strong redistributive connotations and is often linked to the notions of "allocating" and "sharing", kugawira and kugawa. If others benefited from katangale and no one was harmed -except for the state of course -people tended to consider corrupt behaviour as acceptable and morally right.
Of course, these are very general observations and often katangale was also criticised as being harmful and anti-social. Lwanda (2005), for instance, draws a linguistic distinction between katangale and ziphupu. According to Lwanda, the original meaning of katangale refers to the custom of selling food from a basket. He argues that this distinction became blurred between 1999 and 2004 when "neokatangale" and a "culture of we are all corrupt" became entrenched (Lwanda 2005:53). In spite of his criticism of "neo-katangale", he also draws attention to the positive connotations associated with the term katangale used to justify behaviour considered to be corrupt by the law.
Katangale was clearly differentiated from ordinary theft, kuba. Theft was an individual act for one's own benefit. Kuba did not evoke the redistributive associations of katangale. Kuba was always strongly rejected as a wrongful act. Theft was legally and morally wrong and could not be justified as katangale. For example, in the markets it was common practice that a thief who was caught redhanded was subjected to spontaneous mob-justice that often resulted in severe physical punishment or the death of the thief. These observations regarding the concept of katangale indicate that the attitude towards the use of the public office for "private" ends was more complicated and multi-layered than simplistic notions of African "primordialism" suggest.

Corruption and normative pluralism
The moral ambivalence of katangale indicates a more complex normative landscape than the public denouncements of "corruption" suggest. In Malawi, as elsewhere in Africa, certain expectations and obligations arising from kinship relations and other social norms co-exist with the law of the state. Often these are invoked when the transgression of official rules requires justification. Corruption usually is justified in moral terms if it serves altruistic ends as the term katangale indicates. The phenomenon of multiple normative orders co-existing in the same social field or pertaining to the same domain of social life is known as legal pluralism in the legal anthropological literature (Benda-Beckmann and Strijbosch 1986, Griffiths 1986, Melissaris 2004, Merry 1988, Snyder 1993, Woodman 1998. Such a conception of legal complexity harks back to legal sociologist Eugen Ehrlich's (1967) notion of "living law", the adherence to practical norms in a society where the scope of state law is limited.
Studies of the African state also note the existence of multiple normative orders. Olivier de Sardan, for example, points out that "what is considered to be corruption from the perspective of official norms is not, or seldom, viewed in the same light from the perspectives of practical norms and actual practices." (Olivier de Sardan 1999:263). This contradiction is more than a discrepancy between norms and practices but should be analysed in terms of a contradiction between different sets of norms (Médard 1995, Olivier de Sardan 1999.
Of course rules as such are abstract and do not interact by themselves. Rules are instantiated and only conflict with each other through human interaction. Civil servants are very adept in picking their way through the normative plurality. They suffer and profit at the same time from conflicting norms. On the one hand, they often are caught in catch-22 situations since whatever they do will result in a violation of a rule: if you help your brother you might have to violate official regulations but if you do not help him you run the risk of undermining kinship solidarity. On the other hand, civil servants use their position as brokers as a resource always ensuring that they get their share, either by profiting themselves or improving their status and position in their respective networks. The following case study of petty corruption nicely exemplifies the possibilities and constraints offered by normative complexity in Malawi.

A chameleon in action
While conducting fieldwork in Malawi I was able to gather quite comprehensive ethnographic data about corruption in the civil service in spite of the well-known methodological difficulties the anthropology of corruption is confronted with. One of the most complete cases is presented here. It is an account of events that occurred in December 1999 at a civil servant's house where I was staying at the time. The name of the civil servant is fictive to protect his identity and place names, the department or any other information, which might reveal the identities of the people involved, are not disclosed. It concerns a relatively minor transgression of office rules and regulations concerning the use of government vehicles for the transport of deceased civil servants and their family members. The unauthorized and illegal use of government property is known in French as "perruque" and is a very common form of corruption in Africa (cf. Blundo and Olivier de Sardan 2001a).
Mr. Mashanga was a senior civil servant. In his early forties he was at the height of his career. It was amazing to see how he handled his superior officers, subordinates, relatives, clients and all kinds of relationships including that with a European PhD-student who had taken quarters in his compound. Like a chameleon he switched effortlessly between codes and environments, talking a very learned and bookish English at one moment and within a split second putting a subordinate or client at ease with a funny remark in Chichewa. He had contacts everywhere, from the patrons of seedy bottle-stores in town to the highest officials in the civil service. He could arrange almost anything and his subordinates would react on the wink of a finger. Everywhere he was greeted with respect and although he did not belong to the highest echelons of power he could mobilise an impressive network and certainly had a high social status.
He lived in large house provided by the government. His wife, who had also been a civil servant, had passed away a year ago and he lived together with his son and a number of relatives and dependants, some of them in the main house, others in servants' quarters. Two of the younger ones attended school in town. All of them depended on him for accommodation, school fees, food and hoped to find employment through his contacts in the civil service. Mr. Mashanga had enjoyed a good education at a Catholic missionary boarding school and had graduated from the University of Malawi in 1985. His father had been a primary school teacher and had made quite an effort to ensure that his son would get a good education. Mr. Mashanga had joined the service in the late 1980s as a junior professional. Later he had served on different posts where he distinguished himself as able administrator and a good leader who enjoyed both, the respect of his superior officers and the loyalty of his subordinates. Swiftly he had moved up the promotional ladder to the post he held during the events described below.
Mr. Mashanga employed a housekeeper who lived together with his wife, children and mother in one of the servants' quarters on Mr. Mashanga's compound. Although the housekeeper worked in the household and was paid wages it would be inadequate to describe their relationship only in terms of an employment contract. As is common in Africa their relationship could best be described in terms of that between patron and client. They were very close; the man had worked for Mr. Mashanga for more than ten years. One day in December the housekeeper's mother died unexpectedly of malaria. He turned to Mr. Mashanga, his patron or bwana (literally master), for support. This came as no surprise to Mr. Mashanga since he, as bwana, was expected to help his dependants in situations like this. His client had no one else he could turn to because he and his family were poor and the home village far away. The housekeeper was sure Mr. Mashanga would help him because he had a reputation of helping his widespread circle of clients in times of need.
Mr. Mashanga really wanted to help the man who was very distressed and very dear to him, 'like a relative' as he stated. Yet, he faced several difficulties in helping the housekeeper. First of all, he needed vehicles to transport the corpse and the funeral guests to the village where the funeral was supposed to take place. His own car had been grounded for a couple of months. The engine had broken down and the necessary spare part was not available in Malawi and anyway, he could not afford a replacement at the moment since the spare part would cost him at least two months salary. Prices for spare parts had rocketed in the past months due to the high inflation rate of more than 30 per cent. But even if his car would have been running it would not have been really suitable for the transport of the corpse and the family members attending the funeral in the home village. To rent a pick-up truck was out of the question because it was far too expensive for Mr. Mashanga. Instead he turned to his office for help and arranged two government vehicles with drivers.
This entailed the violation of official rules and regulations. According to the Malawi Public Service Regulations (MPSR) (Supplement 1:192.1,192.2) the government provides a coffin and vehicles to transport the body and the guests to the place of burial only for the civil servant him-or herself or a civil servant's family member living at the service station. Despite their apparent rigidity even official regulations leave some leeway since they do not exactly define who qualifies as a family member living at the service station. In a typical civil servant's household, where relatives often stay for weeks or months, this border is often difficult to draw. Stretching the meaning and with a little bit of goodwill even a nephew or cousin on an extended visit could be considered to be a "family member living at the service station" as long as there was some kin relation with the civil servant and the relative died at the service station. In Mr. Mashanga's case the situation was a little bit more complicated. Although his housekeeper had worked for him a long time and was considered by Mr. Mashanga to be a relative, abwale, neither he nor his mother would qualify as family member in terms of the MPSR, which does not recognise fictive kinship.
Mr. Mashanga was well aware of the violation of the MPSR, which he held in high regard. According to him "everything concerning the civil service is in there [i.e. the MPSR]" but in cases such as the funeral of his housekeeper's mother it was necessary to circumvent official regulations that "existed only on paper" as he pointed out. The following day he had arranged two government vehicles with drivers who transported the corpse and the funeral guests from town to the village at approximately 60 km distance. After the funeral in the village the drivers had stayed overnight in a trading centre nearby since it was already too late to return to town. By using government vehicles, Mr. Mashanga was able to keep the costs for his personal burse very low. In the end he only had to pay condolence money which was about 10 percent of his monthly salary, still a considerable sum of money. This money was used by his housekeeper to cover the other costs of the funeral such as food and drinks for the guests. His housekeeper and the funeral guests praised Mr. Mashanga and many expressed their gratitude for his support during and after the funeral as I was told.
Whilst the use of the government vehicles amounted to an arguably minor violation of civil service regulations all involved took utmost care to respect official procedures in order to cover up the illicit use of the cars. Mr. Mashanga submitted a request for the vehicles, the superior officer signed the transport order, the executive officer countersigned it and the drivers filled in their logbooks. With regard to the justification for the allocation of the government vehicles an alternative set of rules was invoked. At the office everybody fully understood that he had to fulfil his social obligations and the impossibility to refuse the request by referring to official regulations. As patron the employer has a responsibility not only for the employees but also for family members. A civil servant denying access to the state's resources would appear selfish and heartless, ubombo. Although employers often complain about this often quite costly obligation, it is virtually impossible for them to reject such a request on their resources since funerals have the highest priority in social life and anybody who appears to have no empathy in such a matter would be threatened by ostracism and witchcraft attacks.
Mr. Mashanga's case exemplifies the moral ambiguity of katangale. The circumvention of official regulations was justified by the moral obligations arising from the relationship with his housekeeper. Brokerage on behalf of his client had consequences for the social networks at the office: he became indebted to his superior officer, the drivers and in a more indirect way to his colleagues at the department who knew about the improper use of the vehicle. Mr. Mashanga's ability to arrange two cars within a day for a trip far away gives testimony to his high status and influential position within the department. Furthermore, considering the ease with which he was able to arrange not only one but two cars with drivers indicates also that there might have existed a debt on the part of his superior officer and other colleagues stemming from earlier acts of katangale committed by them. This web of indebtedness at the office indicates the existence of a certain modus vivendi, a set of unofficial rules providing some guidance in situations where the law in the books is considered too far removed from the predicaments of real life.
The relative ease with which Mr. Mashanga was able to circumvent the restrictions of the civil service regulations does not imply that the official regulations were irrelevant or not applied. On the contrary, they were applied -albeit in a selective manner. As Mr. Mashanga's case shows it can be very easy to circumvent official rules and regulations. In the unfortunate case, however, that a citizen has no relationships with functionaries who could pull a few strings official rules and regulations are usually applied to the very letter -with often Kafkaesque consequences.

Three sets of rules intertwined
The central question that arises from Mr. Mashanga's story is whether his actions were corrupt or not. It could be argued that they were corrupt since he violated official regulations for his personal advantage. On the other hand, it is clear that the people involved considered his actions to be morally right. Therefore we can understand his story as an example of a chameleon's careful manoeuvring between mutlitple normative orders with conflicting rules. Three sets of rules can be distinguished in the case study. First, there are official rules and regulations, then there are the social norms regarding Mr. Mashanga's obligations towards his housekeeper and, last, there is a set of informal and clandestine rules and principles regulating behaviour at the office Mr. Mashanga is able to invoke.

Official rules and regulations
African bureaucrats are famous for their often mind-boggling formalism. Comaroff and Comaroff (2006), for example, comment on the excessive legal formalism and the meticulous care for procedures, stamps and documents in the postcolony, which they describe in terms of a "fetishism of the law". Malawi is no exception in this regard: There is an array of legal instruments regulating the conduct of government officials. The primary code is the MPSR regulating salary scales, conditions of employment, and rights and duties of all government employees. Therefore, they are commonly referred to as "the bible of the civil service" by civil servants. This body of rules is supposed to regulate exhaustively the conditions of service.
The MPSR contains a chapter on conduct and discipline sanctioning for example the use of information for personal gain, the use of public monies or property for private purposes, activities which conflict with the interests of the government or are inconsistent with a civil servant's official duties, etc. There have been only two editions, in 1978 and 1991. Since then the provisions have been altered, updated and supplemented by a vast number of circulars, which are usually issued by the Secretary of the President and the Cabinet and the Secretary of the Human Resource Department. In spite of these amendments the rules, the terms of service and the structure of the civil service have basically remained the same since British colonial rule (cf. Skinner 1963, World Bank 1994. These regulations are supplemented by the penal code prohibiting theft and fraud by civil servants, and in 1995 Parliament enacted the Corrupt Practices Act, a consequence of conditions set by the World Bank and the IMF for their financial support. In addition, numerous circulars of the Secretary of Human Resource Management address specific problems such as the misuse of government vehicles. Several agencies have the task to deal with wrongdoings of public officials. Since colonial times the Auditor General's Office has been the primary department controlling the ministries' accounts. This institution has been joined by the Anti-Corruption Bureau (ACB), the Department of the Public Prosecutor and the Public Accounts Committee of Parliament in the late 1990s.
Thus, state legislation and institutions are not a homogenous system but in itself a rather heterogeneous mix with different genealogies determined by changing geopolitical constellations. From independence in 1964 until 1994 Malawi was under the paternalistic autocratic rule of Kamuzu Banda, since 1971 President for Life, and the MCP, the only legal political party. Kamuzu Banda retained and even strengthened the highly hierarchical and centralised structure of the colonial civil service since it suited his interests. This started to change when a new government under Bakili Muluzi was elected in the first multi-party elections in 1994. The new government had a reformist agenda and implemented a number of reforms to transform the service into a "performance-driven", "transparent" and "accountable" provider of public services controlled by democratic institutions and the general public under guidance of the World Bank, the IMF and bilateral donors. As a result of these demands the government has implemented a civil service reform programme. In Malawi the World Bank adopted the so-called "enclave approach" to governance reform. Hereby new agencies are established that are supposed to function as motors of reform (Dia 1993). So far these reforms have not been very successful in transforming the civil service and structures inherited from the period of British colonial rule uneasily co-exist with newly established agencies with the task to promote institutional reforms.
In spite of the scope of official rules and legislation the actual risk of discovery and punishment for corrupt civil servants is very low. Of course the President and the ministers are eager to point out that corruption is an evil and that no one stands above the law. And indeed in recent years several high-profile scandals involving ministers and high officials have led to temporary arrests, transfers and suspensions although only one conviction. Control by these institutions and the notoriously understaffed Auditor General's Office is incidental at best and enforcement of the various statutes and official regulations seems to be rather restricted to a few isolated cases that are supposed to have a deterrent effect or serve political purposes of political rivals. Furthermore, the rivalry and unclear mandates of these institutions have counter-productive effects hindering rather than enhancing anti-corruption measures.
Under Kamuzu Banda's and the MCP's regime discipline was strictly enforced in the civil service. Most senior civil servants were completely dependent on Kamuzu Banda. He had appointed them to their positions and he had provided them with tobacco estates. The loans for the purchase of the estates were granted by the state-owned Commercial Bank that was also controlled by Kamuzu Banda. Compliance to official rules was further ensured by arbitrary arrests of suspected "confusionists" and "separatists". This "culture of fear", as it was called by my informants, guaranteed compliance to official rules: any misstep could lead to one's dismissal and arrest. The Malawi Young Pioneers (MYP), the paramilitary youth wing of the MCP, had spies and informants everywhere and sudden unexplained wealth would always lead to inquiries. Thus it seems that the reportedly low levels of corruption Malawi used to be famous for prior to the 1990s depended on Kamuzu Banda's system of brutal oppression rather than the rule of law.

Social norms and the ideology of sharing
In Malawi, as in the rest of sub-Sahara Africa, social networks are vital in regard to people's sense of self, status and welfare. A person's status is defined to a large degree by the number of personal relationships he or she entertains and how much of one's wealth is redistributed among kin and clients. Due to the absence of universal and substantial social welfare services people in general have to rely on their networks with kin, colleagues and neighbours for their social security. Most writers emphasise the role of informal or traditional networks with regard to the widespread corruption in Sub-Saharan Africa Olivier de Sardan 2001a, Médard 1982:172;Olivier de Sardan 1999:256, 257). Civil servants act as brokers and patrons within a web of interdependent social networks transcending the sphere of the public office. This position makes them receptive for the temptation of corruption. For many it is almost impossible to refuse the demands of a relative, a friend or an old schoolmate to assist, even if it means they have to circumvent official rules. Of course, not all demands will be treated equally, a relative's request for support might be subject to different considerations than that of a friend and a superior's demand certainly is treated differently than that of one of equal social status.
Generally in Malawi there is a strong moral obligation to redistribute, kugawa or kugawira, i.e. to share, accumulated wealth with kin. The generous patron enjoys considerable status among his clients. In turn, the connection with a powerful patron enhances the social status of a client. The achievement of an individual is always seen as a collective achievement by a potentially wide circle of relatives, which will grow to the same extent as one moves upward the social ladder to the extent that the senior civil servant or the successful businessman will be visited by "relatives" he didn't even know existed. Social control within these networks is usually tight and the implicit threat of witchcraft, ostracism and gossip functions as a sanction inducing conforming behaviour.
The strength of solidarity networks should not be confused with an idealised picture where all corrupt behaviour by public officials is a logical consequence of demands for support. Blundo and Olivier de Sardan (2001b) rightly draw attention to the fact that social or moral norms often merely serve to justify corrupt practices. Civil servants are active agents, brokers who continuously negotiate the terms of their social relationships, making sure that they get their share of the cake. Often civil servants themselves point out that it is no wonder that other officers (of course it were always others) are corrupted when they are confronted by the obligations of extended families and the limitations of low salaries eroded by inflation. This is often not more than an excuse for corrupt behaviour; not everybody is the same: some were not corrupt, others were manipulating and circumventing official rules under severe circumstances only, like Mr. Mashanga, and some were downright corrupt without the pretext of social obligations and solidarity, using the gains from corruption for hedonism and escapism. The numerous civil servants who can be encountered in the bars and night-clubs of the urban and peri-urban areas where they spend a month's salary in one night give ample testimony of this selfish and antisocial aspect of corruption.
Civil servants maintain diverse social relationships with neighbours, colleagues, old schoolmates and fellow members of the church congregation. Especially in an urban environment networks based on residence, social status and church membership are very important. Often these relationships are described in terms of fictive kinship, especially in regard to solidarity. Yet one should be careful to lump all social relationships and personal exchanges under one header as many authors seem to suggest (Ekeh 1975, Médard 1982, Olivier de Sardan 1999, Scott 1979. Civil servants themselves often made a distinction between kin and friends. Of course, relationships with neighbours, colleagues and people from the same congregation can be under strain from demands for support but most civil servants I met stressed the pressure from relatives who expect them to provide support as a major issue of their existence. With regard to friends my informants often stressed that a friend could put less strong claims on their resources. In spite of these complaints their attitude towards expectations by kin is often ambiguous: on the one hand, exaggerated expectations are perceived to be a burden in an environment characterised by severe economic crisis and loss of real income; on the other, the status that comes with the role of a broker and patron is very tempting and kin relations are seen as insurance, a potential asset which has become increasingly important in recent years under the impact of economic crisis.
It would be misleading to represent the rules regulating informal social support as a normative system of defined rights and obligations. It constitutes rather a cluster of basic principles, which order the sphere of personal relationships. Instead of speaking of rights and duties it seems more appropriate to talk about expectations and feelings of obligation. Nevertheless, the latter tend to be felt strongly and people devise the most complicated schemes to avoid or at least to control the demands from their relatives. Exchanges are often never fully consumed to maintain a sense of indebtedness, which leaves open the possibility to ask for favours or services at a later point of time. Parallel, terms and conditions of exchanges tend to be implicit and vague, thus leaving space for changes in interpretation and a continuous process of negotiation. For example, the responsibility of the employer is not limited to his employee. As employer Mr. Mashanga is expected to support his employees and their families in times of distress. The employer is generally expected to arrange for or to pay the funeral and costs of transport of the deceased's body to the place of the funeral, usually the home village. There is no clear rule since it is neither in the interest of the employer nor that of the employee to define more precisely who would benefit from the employer's support. Thus both parties have the possibility to negotiate the conditions of help in the light of the specific circumstances of the situation. A more precise definition would be considered undesirable by both, the employer who wants to avoid to give the impression that he or she is keen to limit his obligations and the employees who have an interest to have as little as possible restrictions on support from the bwana.
This sketch of principles ordering social relationships in Malawi seems to correspond with Ekeh's concept of the "primordial public realm" as opposed to the "civic public" associated with colonial rule, which is devoid of moral imperatives. 19 He seems to suggest that a homogenous set of moral imperatives operates in the "primordial public" and influences the private sphere. In my view, this conceptualisation is misleading: in general the idea of a primordial realm is anachronistic at best and seems to be inspired by a search for authenticity rather than social realities in modern Africa. In present-day Malawi there exists not one universal morality but rather a patchwork of different moralities. Often the differences are negligible, as for example between different ethnic groups who have similar moral imperatives, but sometimes they are considerable. This is certainly the case in regard to Christian morality and kinship morality, which are often at odds and give rise to tension and conflict. Ekeh's concept of the "primordial public" in Africa, which he primarily associates with kinship and ethnicity, ignores the profound influence Christian missionary activities have had on society in Malawi (McCracken 2000, Ross 1996, White 1987. Often civil servants who are devout Christians, especially the rising number of "born-again" or "charismatic" Christians, have high moral standards in regard to 19 Ekeh points out that there exist two public realms in the postcolonial state: on the one hand, the "primordial public" realm associated with traditional kinship relationships, which is "moral and operates on the same moral imperatives as the private realm" and, on the other, the "civic public" of the state's institutions based on the ideology of the colonial administration, which is "amoral and lacks the generalised moral imperatives operative in the private realm and in the primordial public " (1975:92).
an individual's personal conduct. Christian ethics differ in important aspects from kinship-influenced moral imperatives: Charity, for example, is a moral duty for Christians but unlike solidarity between kin with strong social pressure applied by members of the group the obligation is based on the belief in god and the hereafter where good deeds will be rewarded or bad deeds will be punished. For the major Christian denominations corruption and katangale constitute sinful behaviour and for some charismatic churches the devil himself tries to seduce the believers from the path of god by tempting them to act wrongly. This strong rejection of illegal acts, which can be traced back to the acknowledgement of secular power in all Christian churches, 20 differs from kinship solidarity, which does not recognise the distinction between public office and private sphere. Differences between charismatic Christian ethos and kinship morality, however, do not imply that charismatic Christians are less prone to corruption than other denominations or non-believers. I merely suggest here to replace the notion of a "primordial public" by conceptualising a mix of alternative moralities which may have conflicting elements.

The parallel social order in the bureaucracy
All bureaucracies know informal arrangements or codes of conduct at the shop-floor level. In Africa this parallel order appears to be particularly strong and well-developed. Beneath the layer of statutes, regulations and organisational charts a complex web of interpersonal relationships seems to amount to a parallel structure within the state apparatus running mainly along the lines of patron-client relationships. This phenomenon is variously described as "neo-patrimonial state" (Médard 1982), "rhizome state" (Bayart 1993), "shadow state" (Reno 1995(Reno , 2000 or even "criminalization of the state" (Bayart et al. 1999). In Malawi this parallel order does not amount to a "shadow state" in Reno's (1995Reno's ( , 2000 sense who describes a much more institutionalized parallel system in Sierra Leone, where formal structures had been hollowed out completely during the civil war, but a parallel office mores is easily discernible in the Malawian state apparatus.
The parallel social order or office mores is best described in terms of a set of basic moral principles and an unofficial code of conduct based on the principle "respect for the master", ulemu kwa bwana, that guide the interpretation of official regulations and prevail over them in case of conflict. This system soon emerged after independence and constituted one of the main instruments used by Kamuzu Banda and the MCP to ensure absolute control over the state since all civil servants owed loyalty to the party and Kamuzu Banda personally. After the introduction of multiparty democracy the parallel order was transformed into a source of patronage for senior civil servants and politicians who took advantage of the crumbling discipline and the end of one-party rule. Systematic misuse of the public office started among the top cadres of the civil service and spread quickly down to the junior grades who imitated their superiors in an attempt to benefit from the political turmoil of the mid 1990s. Being freed from the constant fear to be disciplined senior officials and politicians of all parties seized their chance to establish themselves as patrons in their 20 Of course there are differences between the various Christian churches. Protestant churches have a more pronounced tradition of being apolitical and recognising secular power than the Catholic Church. Yet, most churches recognise secular laws and adhere to the private/public distinction. own right and appropriated the resources they controlled. In contrast with Kamuzu Banda's system in which only a small circle around the autocrat appropriated the resources of the state one can speak of a democratisation of appropriation with regard to the transition to democracy in the 1990s. The democratisation of appropriation was often seen in the context of the new democratic dispensation and many informants cynically perceived of human rights as a regime that allowed one to exploit the new freedom for personal enrichment or as "the right to grab as much as you can" as one interlocutor wryly put it.
Social relations within the bureaucracy have two central features: asymmetric power relations and the maintenance of indebtedness. Junior civil servants usually drew a sharp line between "the bosses" and themselves who depended on the protection afforded by their superiors. Without the intervention of a patron or friend it was almost impossible to get a salary advance, the permission to attend a workshop, to get a promotion or even to get a job in the first place. For example, in many government departments it was understood that civil servants had to pay a "commission" to their superior officer if they wanted to get a salary advance. Indebtedness was the other important aspect of relationships within the civil service. Mr. Mashanga's story shows how personal relationships at the office produce and reproduce indebtedness thus spawning katangale. The production and maintenance of indebtedness were desirable for both, debtor and creditor. The interest of the creditor was obvious: he or she gained influence over the debtor. Often full return was not in the interest of the creditor who preferred to maintain a degree of leverage on the debtor. But even the debtor often perceives of a debt as an asset. If it is not possible to become a creditor the next best thing is the status of a debtor. To have a debt constitutes a means to establish a social relationship with someone who might develop a sense of obligation towards the weaker party and assume the role of a patron.
The production and reproduction of webs of indebtedness is central to the understanding of the parallel order of social relationships in the civil service. By making other accomplice to corrupt and dubious practices loyalty is ensured. In Mr. Mashanga's case the boss authorised the private use of government vehicles because he owed him a favour or wanted to create a debt. The protection afforded by the bwana is vital for the junior officers and usually people are discharged or prosecuted because of corruption when their master removes his protection. The reasons for such a move by the patron may vary. Often a client is sacrificed in order to stay out of the line of fire when the patron is threatened by investigations, often on the instigation of a rival. Sometimes protection is removed because of violations of the office mores by a client who failed to pay proper respect.
According to the junior civil servants I talked to during my fieldwork "respect for the boss" is the most important virtue in Malawian government institutions. "Respect for the boss" overrules the official regulations, the MPSR. One informant pointed out that "we have laws but we also have by-laws". The superior officer "knows that there is some government regulation but he may divert" whereas the junior civil servant has to obey, "in the office you cannot challenge a senior officer, if you do that he can transfer you" as a junior civil servant put it. My interlocutors often alluded to the office mores as "know how to manoeuvre" or the "rules of the game".
Obviously there exist differences between different departments and sub-units like offices and schools. According to the old adage that the fish stinks from the head, the conduct of superior officers and managers heading a sub-unit usually sets the tone for the rest of the staff. Opportunities for corruption also depend on the functions of a department: the Fisheries Department, for example, differs from a primary school and police officers or accountants tend to have more opportunities than clerks or messengers. Nevertheless there is a set of basic informal rules and codes, which can be found to some extent in all ministries and departments down to the smaller units such as police posts, health posts and schools.
The central principle of "respecting the boss" was only countered by the principle of sharing and allocating resources with dependants and clients mentioned above. The patron who violates this principle often faces silent resistance or even revenge, in the form of corruption allegations for instance. Junior officers tend to tolerate or even condone corruption by the superior officers as long as it is katangale, a system of redistribution in which everybody involved benefits. This attitude of junior civil servants is nicely illustrated in the song Ndiphike Nyemba by Charles Nsaku that was very popular in early 2002. A driver accuses his boss of being selfish, "I am your driver, we left Lilongwe together the previous day coming here to the field to work, there you are with money boss, food you are just eating alone, while your driver ate two days ago". These accusations of being selfish or ubombo are not something to be taken lightly in a society where the obligation to share is a central principle of conduct. In the song Njoka mu udzu by Lucius Banda the junior officer compares his boss to a snake lacking basic human qualities: "You are like a snake in the grass, when passing us by you smile, deep down in your heart you harbour bad feelings towards us, what is that you are missing boss?" He also warns his boss that one day he might have to depend on his subordinates since "life is like a card game, today we are working for you but your children might be employed by us".

Conclusions
"Good governance" policies in Africa are informed by a reductionist culturalism that blames African "primordialism" or "patrimonialism" for corruption and underdevelopment. There is a tendency to represent state institutions as alien transplants "unrooted in local culture" (Dia 1996:30) harking back to notions like the "primordial public realm" (Ekeh 1975). The imagined disconnect between institutions imported by the colonial powers and African society or culture is seen as the cause of "weak states" and "failed states". The ethnographic evidence presented in this article shows that these simplistic dualisms do not account for the normative complexity in Malawi, arguably a stronger state than the usual suspects commonly cited, where several sets of rules co-exist in the same social field. Civil servants in Malawi are well aware of the existence of official rules and regulations governing their conduct and they incessantly develop ways to get around them justifying their transgression with moral obligations and the impracticability of codes existing "only on paper" as Mr. Mashanga pointed out. His case study shows the careful manoeuvring of a chameleon continuously switching normative codes depending on the situation. It is correct that civil servants in Malawi tend to justify corrupt behaviour with the strong expectations of kith and kin to grant access to public resources and Mr. Mashanga's case illustrates this very well but it would be naïve to assume that corruption in Africa is always a consequence of social pressure.
Mr. Mashanga's is a quite harmless and common case of petty corruption but it aptly illustrates the ways corruption works in Malawi. He invoked his moral obligations as a patron towards his dependant to justify his misuse of government property. His superior and colleagues accepted this and condoned the misuse of government property in this specific case thereby abiding to the strong ideology of sharing and caring in Malawi. The idiom of kinship solidarity, however, competes with and is influenced by other moralities such as Christian values. It would be, therefore, misleading to assume the existence of a homogenous African morality. A third set of rules, an informal modus vivendi, exists at the office that facilitates and regulates violations of official rules and regulations that are considered to be too far removed from real life to be always applicable. The existence of a parallel social order at shop-floor level is not peculiar to Malawi or Africa but due to its specific history it has structured corruption in the civil service since 1994.
It is important to remember that civil servants are active agents exploiting ambiguities and gaps resulting from normative pluralism. Even the lowest civil servants occupy a relatively privileged position in relation to the population at large. This makes them important brokers in their social networks and constantly people try to make claims on them by reminding them of their social obligations. Thus, civil servants experience a tension between official rules and their everyday lives. Mr. Mashanga's case illustrates this virtuous playing with different sets of rules to maintain his position both in the office and the "private" sphere.
An analysis that takes indigenous concepts such as katangale and kugawa into account transcends the simple public/private, legal/illegal dichotomies abounding in non-anthropological treatments of the phenomenon. The recognition of normative plurality in the bureaucracy of Malawi reveals alternative dichotomies such as selfish/altruistic and different ideas of conformity and deviance that make it more difficult to believe in the potential of universalistic concepts such as the rule of law and "good governance" to curb corruption. We would be better advised to have a closer look at the alternative sets of rules and moral ideologies for guidance in approaching a phenomenon that is neither inherent to African society nor the symptom of a mere dysfunctionality of an inefficient bureaucratic machine.