Corruption and state-building: The case of Iceland

The Nordic countries are generally considered relatively low-corrupt systems. A host of different factors have been invoked to account for the region’s distinctiveness in this respect, including both institutional and socio-cultural factors. The current study presents an addition to the literature on Nordic state building and corruption through the less studied case of Iceland. It is argued that the Icelandic case is important when weighing the relative impact of socio-cultural vs. institutional factors in the emergence of corruption. Unlike the other Nordic states, Iceland experienced a significant amount of corruption during the twentieth century even if it shares many social and cultural features with the other Nordic states. What makes it stand apart is the lack of a strong institutional basis for bureaucratic autonomy, which contributed to relatively high levels of politicization in public life and greater corruption than commonly found in the Nordic countries.


Introduction 1
According to Transparency International's (2021) Corruption Perception Index, the Nordic countries rank among the least corrupt states in the world, prompting the question which -if any -lessons can be learnt from their experience, with potential applications to other states. How did Denmark become Denmark (Fukuyama 2011;Mungiu-Pippidi 2013)? What was the Swedish road to Sweden (Nistotskaya & D'Arcy 2018;Rothstein & Teorell 2015, Teorell & Rothstein 2015? The Nordic countries, however, have many social, cultural, and political features in common, making the disentanglement of cause and effect difficult. Historically, they share potentially relevant features such as relatively egalitarian agrarian structures, Lutheran state religion, high literacy, a high degree of associational activity, and strong Weberian bureaucracies along with low-corruption politics. The question is how a distinction can be made between coincidental relationships and causality. The present article adds a less studied case, the Icelandic one, to existing knowledge of Nordic state building and corruption. The literature on state-building and corruption in the Nordic countries (e.g. Alapuro 2019;Frisk-Jensen 2014;Papakostas 2001;Rothstein & Teorell 2015) barely mentions the Icelandic experience. A significant amount of clientelism developed in twentieth century Icelandic politics, making it the only Nordic state where corruption was common political practice.
The present paper is a case study of corruption in Iceland based on comparisons with the other Nordic countries. Case studies, according to Gerring (2004, 341) are indepth studies of a "single unit (a relatively bounded phenomenon) where the scholar's aim is to elucidate features of a larger class of similar phenomena". Underlying is the logic of comparative analysis which has influenced important work in political science (e.g., Moore 1966;Skocpol 1979). Comparisons in the present paper are based on Mill's (1874)

method of difference:
If an instance in which the phenomenon under investigation occurs, and an instance in which it does not occur, have every circumstance in common save one, that one occurring only in the former; the circumstance in which alone the two instances differ, is the effect, or the cause, or an indispensable part of the cause, of the phenomenon (280).
The crucial variable of difference in this case between Iceland and the other Nordic countries is that of corruption. Corruption, it is claimed, developed to a greater extent in twentieth century Icelandic politics than in the other Nordic countries. Since the Nordic cases have many other characteristics in common -many of them associated in the literature with the emergence or non-emergence of corruption, differences can tentatively be suggestive of the factors shaping corruption outcomes. These include social and cultural factors as well as institutional ones. The central contention of the present paper is that whereas Iceland broadly shares social and cultural characteristics common to the region, these are unlikely to provide adequate accounts of corruption. Instead, important institutional differences are likely to account for variations in corruption.
Given its limited scope, however, the paper does not attempt to cover all aspects of the occurrence or non-occurrence of corruption in the Nordic countries in the manner which a large-scale comparative historical project might. It focuses on areas where previous research in the other Nordic countries has led to relatively clear hypotheses for which Iceland can be used for further testing. In this respect, the approach is confirmatory, i.e. seeking to verify or falsify pre-existing hypotheses (Gerring 2004, 349).
Thus, Iceland can contribute to the well-known debate between the socio-cultural STJÓRNMÁL & STJÓRNSÝSLA and institutional accounts of corruption (e.g. Jetter & Parmeter 2018). While sharing broad social and cultural similarities with other Nordic countries, important aspects of state-building account for higher levels of corruption during much of the twentieth century in Iceland. The paper starts with an overview of the main features of Icelandic corruption in this period. This is followed by a section comparing social and cultural features of Icelandic society with other Nordic countries, demonstrating an overall pattern of similarity regarding potential corruption-related features. This is followed by a section on state-building. A strong Weberian bureaucracy that prevented corruption in other states was lacking in Iceland, paving the way for clientelist politics.

Clientelism in Iceland
Corruption, in the present context, is defined as "the abuse of authority in which politicians and officials exploit their official position to engage in favoritism, thereby contravening the norm of impartiality in the exercise of authority to obtain direct or indirect personal gain for themselves or persons close to them" (Bergh et al. 2016). Corruption is an umbrella concept for different types of activity, including not only bribes and related activities but clientelism, patronage, particularism, and patrimonialism as well (Rothstein & Varraich 2017). Patronage and clientelism denote "the trade of votes and other types of partisan support in exchange for public decisions with divisible benefits" (Piattoni 2001).
The emerging mass-clientelist parties of the inter-war period in Iceland faced a fledgling bureaucracy and officialdom that had lost some of its previous power position and lacked powerful political allies to protect its position. The subservient position of the bureaucracy made Iceland easy prey for partisan control, with politicians handing out favors on a case-by-case basis. Clientelism offered incentives for joining political parties and created strong expectations of rewards in exchange for political support. It was the foundation on which party organizations were constructed.
Political parties, individual ministers, MPs, or members of local governments served as suppliers in the Icelandic patronage system, often operating through a network of contacts across different levels. Parties were strongly office seeking to gain access to the benefits from control over the executive (Indridason 2005). Recipients of the spoils included businesses, localities, and private individuals. Although a detailed description of clientelism in Iceland is beyond the scope of this paper its main features are well known and commonly recognized. Its main features were probably associated with favoritism rather than bribes or extortion. In the case of businesses, the main contenders were the co-operatives, supported by the rural Progressive Party (PP), and private enterprises, represented by the Independence Party (IP). Competition between the two could be tough; however, in some cases, they agreed upon a system that would lead to a sharing of patronage opportunities, generally referred to as the "halving-system" (helmingaskiptakerfi). This was associated with extensive political interference in the economy, ranging from direct state control over most of the financial sector to extensive case-by-case interference in key economic sectors, contracts, and procurement, in addition to the state regulation of foreign trade (Kristinsson 1993(Kristinsson , 1996; on foreign trade specifically, see Jónsson 2017). The media-party controlled until the 1970s-was mostly silent on this aspect of Icelandic politics. A significant role of MPs was to facilitate the provision of investment and infrastructural development for their constituencies. Pork-barrel politics, especially in the more rural constituencies, was a measure of how successful a representative was (and still is to some extent, see Hlynsdóttir & Önnudóttir 2018). Local governments also played a role in bringing home the bacon and distributing it locally-something more commonly found in Southern Europe than in the Nordic countries (Kristinsson 2015).
Individuals associated with the main parties drew various benefits from party connections, including preferential treatment regarding housing, loans, and welfare. The most notorious use of individual level patronage, however, developed in relation to public appointments, ranging from unskilled manual jobs to the highest administrative positions. People with weak qualifications were often appointed to senior positions, as the education of top officials during the twentieth century reflects.  (Kristinsson 2021) According to Table 2, the proportion of university graduates fell from 92% in 1917 to 61% in 1949-a very low figure in comparison with other Western systems (Aberbach et al. 1981). With the gradual retreat of patronage, the administration became increasingly professionalized; however, it was only toward the end of the century that levels of education similar to those prevailing at the beginning were attained.

STJÓRNMÁL & STJÓRNSÝSLA
Enforcement is a problem in any clientelist system. What suppliers want is political support in exchange for favors. In smaller communities, the delivery of political support could be monitored informally; however, in larger ones, more elaborate measures were called for. For example, the IP in Reykjavik operated an extensive system of "confidantes" who reported the political views of their neighbors and colleagues to the party headquarters, covering 120 districts of the city and every firm with over 10 employees (Jóhannesson 2010, 260).
With growing professionalism and recognition of the proper norms of administration, clientelist practices met increasing resistance from the 1960s onward. Since the 1980s, a series of reforms have undermined it significantly.
While hard data illustrating variations in clientelism in the Nordic region is difficult to come by, the Varieties of Democracy (V-Dem) project offers useful insights. Its index of clientelism, measured on a scale from 0 (low) to 1 (high), is presented in Figure 1. Note: The index is based on indicators for vote-buying, particularistic vs. public goods and programmatic vs. clientelistic party linkages. See Coppedge et al. (2022). Data obtained at https://www.v-dem.net/ The data reflects different evaluations by country experts concerning the extent of clientelism in the five countries. Iceland stands out, with a much higher level of clientelism than the others, while Denmark and Sweden have hardly any experience of the phenomenon. Norway joins them in the post-war period while Finland most of the time lies between Iceland and the others. Clientelism in Iceland reached its zenith during the midtwentieth century but has receded after that to a level more like the other four countries.

The social and cultural context
In large comparative studies, some of the correlates of corruption include social and cultural features that are seen as the facilitators of or hindrances to corruption (e.g. Börzel & van Hüllen 2014;Dimant & Tosato 2018;Treisman 2000). Before delving into the particulars of Icelandic state-building, it is important to establish that social and cultural features were, in fact, broadly similar to the other Nordic countries. The aim is not to present a general evaluation of social or cultural differences between Iceland and the others -but to focus on factors which according to prior research may be considered relevant in the present context.
International comparisons indicate that wealth and development levels are strong correlates of low corruption (Treisman 2015). This does not answer how the direction of causality flows, since corruption may both impede economic growth and reflect a general state of poverty. In the present context we are primarily interested in the causes of corruption, rather than its consequences, and the important question is thus which mechanisms associated with wealth might help prevent corruption. Treisman (2015, 440) suggests that economic development reduces corruption "through the rationalization of public and private roles and the spread of education, which renders abuses harder to conceal". Economic development started late in Iceland compared to Scandinavia, but by international standards it was not a poor country for much of the twentieth century and with relatively high levels of education. While further research may reveal more on this subject the present state of knowledge does not suggest that poverty played a major role. 2

Religion
Protestantism is generally associated with lower levels of corruption, although it seems that being protestant in a historical sense is more important for present day corruption than current religious tendencies (Gokcekus 2008;Mensah 2014;North et al. 2013). While the way religion affects corruption is not a major concern of the present paper, the usual suspects are cultural and institutional influences. Hien (2018, 82) argues that "the predestination model embodied in Catholicism (in contrast to Protestant and reformed Protestant models) has some inbuilt features, like forgiveness, indulgences and repentance, that make fudging morally acceptable". At the same time, however, the state-church conflict in several countries (e.g., Italy) may have undermined the legitimacy of the state and weakened its ability to establish a coherent and impartial administration. By contrast, the close association between the Lutheran protestant church and the Nordic states provided the latter with a powerful tool for penetrating society, obtaining information, and controlling the lives of ordinary citizens, in addition to legitimating state power (Nistotskaya & D'Arcy 2018).
After some initial resistance, the Icelandic reformation was firmly in place by 1550 with the execution of the country's last Catholic bishop. Religious freedom was not formally restored until 1874 by Iceland's new constitution; however, the Lutheran state church remained the predominant religion, observed by close to 100% of the popula-

STJÓRNMÁL
& STJÓRNSÝSLA tion well into the twentieth century (Statistics Iceland 1920). As late as the early 1970s, 95% of the population belonged to the Lutheran state church, while only 1% was governed by other religious communities (Statistics Iceland 1984). Although the position of the state church has weakened in the last few decades, it still enjoys hold over approximately two-thirds of the population.

Agrarian structure, family patterns, and nepotism
Uslaner claims that economic inequality and cultural factors lie at the root of corruption, while institutional accounts of corruption fail to consider the effects of inequality on generalized trust (Uslaner 2009). Inequality creates patterns of particularized trust by dividing people into patrons and clients. Much earlier, Banfield suggested that "amoral familism" is detrimental to the development of social capital (Banfield 1958; see also Putnam 1992). Agrarian structures in the Nordic region were relatively egalitarian during the crucial phase of state formation, promoting the formation of state structures that depended more on generalized norms than particularistic networks of benefactors and beneficiaries (Alapuro 2019;Jensen 2018;Papakostas 2012). Furthermore, family structures in the Nordic region were based less on villages or extended families than in Southern Europe and more on individual farms and nuclear families. Migratory patterns were less family-centric and more individual-based, leading to less reliance on nepotism (Papakostas 2001). A politically weak aristocracy meant that the emerging bureaucratic state had to deal with neither the Roman Catholic Church nor a domestically powerful aristocracy as strong competitors for power and privilege.
Iceland, in all of this, conforms broadly with the Nordic pattern rather than South European ones. While considerable inequalities existed in the pre-market agrarian economy, they were modest by international or even Nordic standards (Karlsson 2008). Landed nobility never existed in Iceland and agriculture was based on scattered family sized farms. No villages existed except for a few merchant stations along the coast. Industrialization began in the fisheries in the early twentieth century; however, employment in the emerging urban economy was mostly not migratory but based on permanent settlement.

Literacy and education
Education levels seem to hold considerable predictive value for corruption; however, as Uslaner and Rothstein highlight, there exists an important historical lag (Uslaner & Rothstein 2016). Educational levels in 1870 predict corruption levels in 2010 rather well. Higher levels of education facilitated equal citizenship and reduced people's reliance on clientelistic practices. Promoting literacy served state-building and the creation of citizens with strong national identities. The emphasis on individual salvation in Protestantism and bible reading facilitated literacy, with the double benefit of high levels of economic growth (North et al. 2013) and low levels of corruption. Educational levels in the Nordic countries in the second half of the nineteenth century were among the highest in the world.

STJÓRNMÁL & STJÓRNSÝSLA
Corruption and state-building: The case of Iceland Iceland basically followed the Nordic pattern. Guttormsson (1988) argues that the spread of written texts in the vernacular after the reformation in Iceland went hand in hand with the spread of literacy after the late sixteenth century. The clergy played a key role in introducing full literacy and reading aloud daily from the scriptures was mandatory in every household. Around the mid-eighteenth century, over one-third of the grown population had basic literacy skills, and by 1800, illiteracy had probably become rare or exceptional (Jónsson 2020). Control over reading material was a powerful tool of social control, although indicators of growing interest in a broader range of texts apart from religious ones or the Sagas could be observed in the late nineteenth century (Gunnlaugsson 1991). Iceland, like the other Nordics, entered the twentieth century with a largely God-fearing and literate population.

Associations and social capital
According to republican political thought, good government rests on "a virtuous, enlightened, and engaged citizenry." Mungiu-Pippidi (2015, 105) finds a strong and significant relationship between the number of associations per capita and control of corruption (Mungiu-Pippidi 2015, 107-108). This, according to her, suggests that building sustainable collective action networks may help solve the collective action problems that often prevent effective control of corruption.
High levels of associational activity are often cited as a facilitator of the development of low-corruption mass politics in the Nordic countries (Alapuro 2018). According to Rothstein (2014), social movements strengthened the collective, as opposed to particularistic, conception of personal interest by virtue of their inclusiveness and nondiscriminatory nature. Such movements included the religious layman movements, free churches, the temperance movement, women's groups, and the emerging labor movement.
In Iceland, associational membership was mainly confined to the administrative elite during the first half of the nineteenth century; however, a great variety of associations had emerged toward the end of the century (Lárusson 2021;Róbertsdóttir 1991). A dense network of associations began to emerge during the last decades of the nineteenth century, including educational and religious organizations, social clubs, co-operatives, trade unions and women's organizations (Jónsson 2001). Although no quantitative estimates of the spread of organizational membership in the early twentieth century exist, it was undoubtedly common. Membership of the co-operative movement was widespread, especially in the rural communities and by the Second World War, the co-operatives claimed (possibly exaggerating) to have members in almost 40% of the households (Ásgeirsson 1988). The labor movement had organized 25% of the working class by 1930 and approximately 80% by the mid-century. The present-day union membership in Iceland is proportionally the highest in the world (Aðalsteinsson & Guðlaugsson 2019), while-apart from the unions-over 60% of the adult population claim to be members of additional associations (Kristinsson 2001).

State-building
At the beginning of the twentieth century, Icelandic society was not only politically a part of the Nordic region, through its association with Denmark, but was also socially and culturally an integrated part of it. As indicated above, this was reflected in religious orientation, social structures, high levels of literacy, and widespread associational membership. There might be other, still unaccounted for, factors affecting the difference in corruption levels; however, the above include the ones most closely associated with current theorizing. Size, for example, is not a strong correlate of corruption. With less than 400 thousand inhabitants, Iceland is a very small state-even in comparison with the other Nordic states-but corruption does not discriminate based on size. It equally affects small groups, village communities, medium-sized states, and large empires. If size matters, it does so through the medium of other variables. 3 The critical difference regarding corruption between Iceland and the other Nordics is the pattern of state-building. The experience of modern democracies suggests that three factors play a significant role in the development of low-corruption states. First, such states need rulers dedicated to the development of an efficient and non-corrupt administration. Since rulers are often among the main beneficiaries of corruption, theories of corruption need to provide a satisfactory account of why state-builders might find a clean administration to their advantage and where they might find allies for such policies (Shefter 1977). Second, a properly functioning Weberian bureaucracy is a sine qua non of low corruption. Managing the transition from particularistic government to a bureaucratic one requires several organizational features contributing to the establishment of impartiality as the guiding principle of administrative procedure. Finally, an effective barrier against case-by-case political intervention into administrative tasks must emerge before the crucial handover of executive power to elected politicians. Bureaucratic autonomy must precede the crucial steps toward electoral democracy (Shefter 1994).
Shefter's thesis is tested for Iceland in the present paper. His approach is that of historical institutionalism, suggesting that temporal sequences and critical junctures matter. The timing of the crucial moves towards democratization of the executive relative to bureaucratization determined a system's propensity to clientelism. To support Shefter's theory we must show that bureaucratization of the Icelandic administration was relatively underdeveloped at the introduction of parliamentary government in 1904 which contributed to the subsequent development of clientelism.

The motives of state-builders
As the main principle of administrative procedure, impartiality is associated with common ideas of justice, but is often shunned by powerful elites. Many states are governed by self-seeking elite groups that thrive on corruption. These are unlikely to initiate largescale operations to minimize corruption, except under pressure. Historically, external vulnerability has been the pressure point for developing a bureaucratic and efficient administration (Huntington 1968). Even advocates of civil society-driven good governance admit that recent examples are few (Mungiu-Pippidi 2013).

STJÓRNMÁL & STJÓRNSÝSLA
Corruption and state-building: The case of Iceland Corruption is difficult to control without a dedicated effort by state-building elites. Various types of corruption were common in the Nordic region during the early modern period (Jensen 2018;Teorell & Rothstein 2015). The incentives for strengthening state capacity and universal methods of administration varied, but also had something in common, i.e., external vulnerability.
Two states emerged out of the Kalmar Union that united the Nordic region between 1397 and 1523. Denmark originally included Norway and Iceland, while Sweden included Finland. The two states were regularly at war, and both suffered heavily during the Napoleonic wars, resulting in Denmark losing Norway to Sweden (1814Sweden ( -1905 and Sweden losing Finland to Russia (1809Russia ( -1917. Simply ensuring the survival of their states and forms of government was a major incentive to fight corruption for the statebuilding elites in Denmark and Sweden (Jensen 2018;Teorell & Rothstein 2015). Teorell and Rothstein (2015) suggest that it is not just preparation for wars that makes states, but how actors come to terms with losing them.
The situation in Finland and Norway was different. The former was a 'Grand Duchy' within the Russian empire, while the latter had its own constitution in a union with Sweden. Both enjoyed a relatively autonomous status in many ways, although their ruling elites had to tread a fine line in their relationship to the controlling state. This was achieved by building domestic alliances around the core administrative elites (both states have been described as civil servant states) while maintaining, for the most part, cordial relations with the Swedish/Russian authorities (Alapuro 2019). In Norway, the bureaucracy could depend on the constitution as well as a relatively democratic political structure (including the Storting) in the nineteenth century, which helped the country mediate between the domestic population and the Swedish state. In Finland, the bureaucracy depended more on bi-lateral relations with Russia in the last decades of the nineteenth century, seeking political loyalty from the Finnish-speaking majority while itself remaining loyal to the Russian state. In each case, however, the "civil servant state" required legitimacy based on impartiality to play a state-building role (Alapuro 2019).
The circumstances of Icelandic state-building were more problematic. Parliament, the Alþingi, played a key role in Icelandic state-building, while the domestic civil service-such that existed-either opposed steps toward independent statehood or played second fiddle. While parts of the civil service were involved with legislative reforms of various kinds, authoritative accounts of Icelandic state-building (such as Karlsson 2008) seem largely to equate state-building with the struggle for independence, with the bureaucracy -at best -in a marginal role. The Icelandic independence movement, which began toward the mid-nineteenth century, was only inspired by liberal enlightenment ideas to a small extent. Even among the intellectuals, supporting the movement, romantic nationalism, and glorified versions of a golden past were a major influence, while its political basis among peasant farmers has been described as reactionary and inwardlooking, reflecting fears against liberal principles of Danish constitutionalism (Hálfdanarson 2001). 4 Such voices were influential in the Alþingi, which was established in 1845 as one of several consultative assemblies in the (still) absolutist state of Denmark.

STJÓRNMÁL & STJÓRNSÝSLA
The civil service in Iceland appertained to the Danish ministries in much the same manner as any Danish region until 1872, apart from a special treasurer (landfógeti) (who, nonetheless, had a counterpart, e.g., in the Faroe Islands). In the period 1872-4, however, the Danish authorities re-organized the Icelandic system, giving limited legislative powers to the Alþingi-a constitution which was in many ways a miniature version of the Danish one-instating a Governor General to head the administration under the supervision of the Iceland ministry in Copenhagen. While the Governors General were all of Icelandic descent, they were a part of the Danish administration, and the Iceland ministry was served by the Minister of Justice in the Danish government. The Governors were generally considered the representatives of foreign rule in the country and a conservative force on the issue of independence. Their close association with the Danish government probably played a role in limiting their political influence.
Hálfdanarson suggests that the failed National Assembly of 1851 -initiated to clarify Iceland's constitutional status -played a critical role in subjugating the officials.
… the National Assembly ended for the most part the active political leadership of civil servants in Iceland. Until then, government employees in Iceland had always regarded themselves as the worthiest outward representatives of the country. … The officials were therefore … among those most strongly advocating the re-establishment of the Alþingi in the 1830s and 1840s and they were also prominent among the leaders of the movement preparing for the National Assembly. … In the aftermath of the National Assembly the Government put an end to such improprieties and demanded of its officials that they make up their mind about which master to serve. For most, the choice was easy and few proved willing to abandon advancement and secure incomes which the administrative system offered. Their place was partly filled by intellectuals from without the civil service, students, scholars, and lawyers, who took over as contacts between the independence movement and the Government … (Hálfdanarson 1993, 32).
With the introduction of Home Rule in 1904, when leaving office, the last Governor gave the following account of the dilemmas he faced: … I always felt the Governor General was -so to say -like a louse between two nails -the nails in the Alþingi and the nails in the government … It is not enough to be … a passable administrative official, to be able to deal adequately with daily routines. More is required of the Governor General, especially when the Minister is Danish, unfamiliar with the country, unfamiliar with its ailments and needs, and therefore incapable of finding the correct measures to mend them. I have experienced this strongly, but what I lacked is initiative, creative ideals and other talents for paving new ways for progress. I have not been in

STJÓRNMÁL & STJÓRNSÝSLA
Corruption and state-building: The case of Iceland favor of great changes, probably too little, I have been conservative, probably too conservative. I hope that I have not done any harm, directly, in my conduct of office, although considerable indirect damage by not doing what could be expected of me (Stephensen 1904).
Members of the Governor General's group formed the core of the Home Rule Party, which came into power in 1904 with the introduction of parliamentary government and Home Rule. It comprised an alliance of establishment politicians and urban interests in a similar manner as Scandinavian conservative parties (cf. Lipset & Rokkan 1967, 37). It had two main weaknesses. First, it was conservative regarding changing Iceland's relationship with Denmark, which had enormous mobilizing power among voters. Whenever nationalist sentiments flared up, the party suffered electorally. Second, the electoral system was highly skewed in favor of the rural and peripheral areas where the Home Rule Party was relatively weak. Its votes weighed less than those of its opponents. First, state-building had to rest on a legal foundation such as the rule of law and civil service legislation. Second, it required an organizational structure following the main principles of Weberian bureaucracy. Third, it needed a professional civil service based on the principles of meritocracy. The development of these core components in Iceland was clearly more problematic than in the other Nordic countries.

Rule of law
If impartiality forms the basis of a non-corrupt government, the universal application of law must be its prerequisite. Systematization of legislation and meritocratic recruitment to the administration were established in Scandinavia during the nineteenth century (Jensen 2018;Rothstein & Teorell 2015). Iceland lagged. Unlike Denmark and Norway, the introduction of absolutism in Iceland in 1662 was not followed by a systematic codification of existing law. Icelandic law remained separate from Danish law and, by comparison, relatively unsystematic. This led to prolonged legal uncertainty as various sources of law-some of questionable validity-were applied (Björgvinsson 1990). Moreover, the judiciary-with the Alþingi as the domestic court of appeal-was in a state of disarray until it was abolished and the district court (landsyfirrétturinn) was established in 1800, but even then, judicial practice remained unsystematic and inefficient (Agnarsdóttir 2008). At the county level, there was no separation of functions as the county sheriffs were heads of police, prosecutors, and judges-an allin-one system that remained in effect until 1992, when it was found to conflict with the provisions of the European Convention on Human Rights

STJÓRNMÁL & STJÓRNSÝSLA
Despite efforts to strengthen the rule of law in Iceland around 1800, in practice, it suffered well into the twentieth century partly due to the lack of sufficient dedication from the executive branch, and partly because of the absence of university level education in Icelandic law (until 1908) and an Icelandic supreme court (until 1920) for authoritative interpretation and guidance. The independence of the judiciary was established in the constitution of 1874; however, the exposition of the judiciary has remained sketchy and less elaborate than in even the Danish constitution of 1849. Throughout the twentieth century, judges were appointed by the Minister of Justice in Iceland, whose freedom of appointment concerning their basic qualifications was restricted by law but often influenced by partisan considerations.
Civil service legislation is a crucial feature of the legal framework for an impartial administration. From a comparative perspective, the existence of civil service laws was a critical factor preventing the access of politicians to patronage (Epstein 1980). Largescale patronage can be prevented "if governmental agencies are protected by civil service statutes and other general laws that specify how public benefits and burdens are to be distributed and that thereby prevent politicians from intervening in the administrative process on a case-by-case basis" (Shefter 1994, 411). Hence, the lack of civil service legislation in Iceland upon the establishment of Home Rule was a critical factor contributing to patronage. Administrative rules were rudimentary and often based on convention, while, in some cases, Danish legislation was used for guidance but had no formal status after 1904 (Jóhannesson 1974, 17).
General Civil service law in Iceland -comparable with the Norwegian and Danish acts of 1918 and 1919 -was enacted in 1954. The Scandinavian states, on the other hand, had de facto established a meritocratic administration during the nineteenth century. The Icelandic legislation during the first half of the twentieth century was fragmented and primitive (Jóhannesson 1974, 98) and political practice favored political appointments. Other legislation concerning the public sector (e.g. on administrative procedure and the right to information) also lagged.

Weberian bureaucracy
Although organizational principles in central administration varied between the West-Nordic and East-Nordic countries, the principles of bureaucratic organization, such as formal procedure, a clear division of labor, hierarchisation, and impersonal application of rules were well in place in the Nordic countries by the beginning of the twentieth century (cf. Herlitz 1958Herlitz -1963.
An Icelandic civil service emerged only with Home Rule in 1904, starting from scratch, with hardly any experienced high administrators, formal rules, or traditions to build on. The ministry, which was established in Reykjavik, had only a handful of employees, including the Minister and a simple structure dividing the offices into I, II, and III bureaux. The offices of the governor general, district prefects, and treasurer-the main pillars of the civil service-were abolished. Only one minister served at a time in 1904-1917 as the highest official of the ministry, and given the small scale and lack of procedure, he was very much present in the day-by-day running of the operation.

STJÓRNMÁL & STJÓRNSÝSLA
Corruption and state-building: The case of Iceland The number of ministers increased to three in 1917; however, coalition government was accompanied by neither hierarchy nor collegial decision making in the cabinet. The organization of the ministries was rather unclear, and in some cases, it was even debated which ministries actually existed. Ministerial government became the norm of cabinet management, making ministerial autonomy the main principle of executive governance. Agencies fell under the authority of the ministries but as they grew in numbers, they were increasingly given advisory boards that functioned as watchdog mechanisms (Kristinsson 1993). A large number of boards and committees remains a strong feature of the Icelandic administration compared with the other Nordic countries (Óskarsdóttir 2018). Nowhere, however, was the symbiosis between politics and administration as close as in local government (Hlynsdóttir 2015). With many municipalities-most of them very small-administrative professionalism at the local level was low, and while there existed supervisory authorities at both the county (sýslunefndir until 1987) and ministry level (eftirlitsmaður sveitarfélaga until 1962), these were professionally weak themselves and played a small role. Many of the Icelandic mayors were, in fact, elected political leaders and chief administrators at the same time.

Professionalization of the civil service
Broadly speaking, the Nordic civil service became professionalized during the nineteenth century. In Sweden, what remained of aristocratic privileges and the accord system were abolished during the second half of the century (Rothstein & Teorell 2015). In this respect, Iceland followed the others to a degree, although nepotism remained common. However, with the onslaught of clitentelism in the twentieth century, setbacks occurred regarding the professional qualifications of civil servants (cf. table 2 above), as appointments became more politicized. While the impact of politicization in other Nordic countries is sometimes debated, they are unlikely to have experienced anything approaching the Icelandic level (Erlingsson & Kristinsson 2020).
Leadership of the civil service in Iceland remained in Danish hands until 1904. Lower-level officials, such as sýslumenn and hreppstjórar, however, were, as a rule, Icelanders, and, overall, the legal education of sýslumenn was comparatively good (Gustafsson 1985). Icelanders began appearing in higher-level positions from the mid-eighteenth century, except for the highest-level administrative post, that of the higher-ranking prefect who was almost always Danish. Of the three Governors General serving in 1873-1904, the first was Danish (although of Icelandic descent) 5 , while the second two were Icelanders (from 1882). Icelanders at the higher levels of the civil service generally had a background in legal studies in addition to some work experience in Copenhagen. Since such positions were generally not advertised, appointments depended on connections to an important degree, and the Icelandic part of the civil service was highly influenced by family networks. Distance and difficult communications made it difficult for the authorities in Copenhagen to effectively control its Icelandic outpost, and in many cases, the authorities in Copenhagen had little confidence in their representatives in Iceland. Misconduct was considered common, and the only Icelander to occupy the post of STJÓRNMÁL & STJÓRNSÝSLA high prefect was effectively suspended in 1803. Hreinsson suggests that the Icelandic elite formed a coherent network based on kinship and marital ties, controlling key positions in the administration (Hreinsson 2003). Despite relatively high levels of education among this group, however, its power basis was more traditional than bureaucratic. It was not supported by an Icelandic university until 1911, which could assume intellectual leadership, like what happened in Norway and Finland. Until then, the University of Copenhagen served as the university for Iceland.
Lower-level officials were probably not closely watched by their superiors, and the county sheriffs, in fact, rented their offices (Gustafsson 1985). In 1877, they were put on permanent salaries and obliged to pay taxes, but they continued to enjoy various side benefits and privileges until the 1980s (Kristjánsson 1987). Clergymen were appointed by the bishop, while in 1886, an act of parliament introduced the direct elections of clergymen by congregations, making the process far less dependent on the church hierarchy than before. There never existed an Icelandic military of any kind, which may have contributed to the relatively formless nature of the emerging Icelandic state.
Positions in the civil service were in great demand, partly for the salaries but also because of side benefits and prestige associated with these posts. The basic rule of lifelong tenure was adopted from the Danish constitution, and officials earned the right to a relatively generous pension-a privilege probably much resented by the general population (e.g. Alþingi 1916).
Missing from the Icelandic administration of the nineteenth century was not so much professional qualifications as effective leadership, rules, and organization which could bolster its capacity for collective action. The loyalties of public officials may have been stronger to the networks to which they belonged than the formal hierarchy. As a social group, they represented privilege, although less exclusive by the end of the nineteenth century than at its beginning. In the Icelandic political landscape, they represented privilege more than national leadership.

Democracy and bureaucracy
In an autocratic regime, the bureaucracy may control sufficient resources to guarantee its power and autonomy. Control over the means of violence, legitimation, and priorities of public policy may suffice to maintain a stable regime. However, with the introduction of democracy, bureaucratic elites will need powerful allies to keep the administrative system intact. Every politician, upon coming into power, is faced with the temptation of using the administrative system selectively as an instrument of control. This is where constitutional government, rule of law, an independent judiciary and a well-functioning bureaucracy become essential for maintaining bureaucratic autonomy (Lundquist 1997).
Bureaucratic autonomy is an important condition for non-corrupt politics (Shefter 1994). Its components may include the broad features of constitutional government (Rothstein 2011), or more narrowly, meritocracy in the administration (Dahlström et al. 2012). However, neither constitutional government nor meritocracy enforce themselves automatically. According to Shefter (1977), the essential factor preventing corruption

STJÓRNMÁL & STJÓRNSÝSLA
Corruption and state-building: The case of Iceland in the wake of democratization is the development of a "constituency for bureaucratic autonomy" consisting of leaders who wish to avoid the development of patronage. "Leaders of a constituency for bureaucratic autonomy … can broaden their basis of support by entering into a coalition with groups that seek through general rules to obtain privileged access to public offices and benefits" (Shefter 1977, 413). Historically, according to Shefter, primarily two types of coalitions existed, namely absolutist coalitions ("creating a modern, centralized, bureaucratic state to replace the decentralized standestaat of the earlier era") and progressive coalitions ("where the role played by absolute monarchs in the first coalition was played … by a rationalizing middle class, whose ideology was Benthamite in England, Positivist in the continent, and Progressive in the United States"). Although the leading elements varied in the two types of coalitions, they were socially not dissimilar. "In order for a rationalizing bourgeoisie to bring about administrative reforms and thereafter to be successful in providing political backing for a reformed civil service, it had to draw support from elements of the groups-be they aristocrats, patricians, or less exalted public servants -that held positions under the earlier regime" (Shefter 1977, 416). Iceland's closest equivalent to a constituency for bureaucratic autonomy might be considered the Home Rule Party and its successors. Its core group consisted of civil servants and the remnants of the old Governor's group in the Alþingi. Its electoral basis was strongest in the towns and more prosperous rural areas (Kristinsson 1991). But it operated in a system lacking the main features of bureaucratic government. Its leadership was from the start criticized for nepotism and cronyism in public appointments. This tendency reflected earlier practices, where the domestic elite disposed of positions on particularistic grounds, even if not all positions were handed out in this manner, as the official historian of the Ministry in Reykjavik is careful to point out (Jónsson 1969). At this point, political appointments were not made to feed party organizations for the simple reasons that general membership organizations on the ground did not exist until the inter-war period. They should more properly be seen as attempts by prevalent elites to bolster their networks.
The 1910s and 1920s saw the entrance of new elites on the political scene that largely replaced the old patrimonial system at the center of power. The greatest threat to the established elite emerged in 1916 with the formation of the Progressive Party-a farmers' party formed to safeguard the interests of farmers and the rural areas in the face of growing urbanization. Several factors worked to the advantage of the PP, making it the main party of government in the inter-war period. First, the effective solution of the independence issue with full sovereignty in 1918 put the old political forces in disarray and created an opening for a new party based on new issues. Second, vast disproportionalities in the electoral system awarded rural votes a large advantage in parliament. Finally, the new party made use of new organizational techniques that led to the creation of an alternative network to that of the established elites, combining the rural co-operative movement, the party organ Tíminn (where effective leadership lay), and a relatively centralized party organization (although a formal membership organization was not estab-

STJÓRNMÁL & STJÓRNSÝSLA
lished until the 1930s). Political appointments became a political strategy, especially after the party was able to form a single party government in 1927, thereby undermining and threatening the position of the old elites. Control over public appointments became a central feature of political competition between the established power groups and their challengers in the Progressive Party.
The forces on the right gradually re-organized to meet the threat from the advancing class parties (a smaller Social Democratic Party (SDP) had been formed in 1916 as well) and finally merged in the Independence Party in 1929. Within the IP, a new elite breed of entrepreneurs rose to power, sidelining, to an important degree, the remnants of the nineteenth century establishment. While this was, from the beginning, by far the largest party in electoral terms, the electoral system negatively affected its parliamentary strength, and the PP retained its central position in terms of government leadership until 1959, when a change in the electoral system reduced its advantage. From the start, the IP was on the lookout for new organizational techniques but ended up copying many of the features that had served the PP so well. Thus, by the 1930s, both the major parties had adopted membership organizations as part of an extensive patronage system that gradually reached most spheres of society. In contrast, the working-class parties, the SDP and the Communist Party (CP, formed in 1930), were relative outsiders. The SDP, nonetheless, is unlikely to have remained excluded from patronage for very long.
The membership organizations of the political parties developed gradually in the inter-war period but resembled in some ways more the looser American parties described by Epstein than Duverger's (1964) European mass parties (Epstein 1980). The PP and IP were internally created parties, and the parliamentary groups played a dominant role within the organizations along with the party leaders. While the SDP and CP were externally created, they never achieved organizational solidification comparable with the Nordic social democratic parties. In effect, their organizations functioned in a manner not much different from the other parties. All the parties developed a formal membership organization in the period 1916-1930, which served as a tool for campaigning and contacting supporters. In effect, the organizations were rather loose, and a clear distinction between members and supporters was not maintained in most cases. Membership fees were not systematically collected, at least in the larger IP and PP, and all the parties relied on contributions from organizations, businesses, or a few well-to-do individuals to a great extent. All the Icelandic parties were strongly office-seeking -in search for material benefits -and the competition for government position was strong. The parties, however, were flexible with regard to government policies, and the coalition game was open in the sense that all parties were potential coalition partners. Policy emphases were flexible and tended to vary between elections. Comparisons between the election manifestos of the Icelandic PP and that of the Norwegian and Swedish farmers' parties show that they were issued in a more irregular manner. They were written by different party organs and varied substantially more in their issue emphases (Kristinsson 1991). Their purpose was to win elections and gain access to power more than to present clear policy alternatives to supporters.

STJÓRNMÁL & STJÓRNSÝSLA
Corruption and state-building: The case of Iceland The party networks were cast wide and deep, permeating almost every section of the population. The political elites were the key elites with close contacts to public administration, businesses, interest groups, media, and the academy and culture to some extent as well (Kristinsson 2021). At its high point, around the mid-twentieth century, the parties were in control of most spheres of Iceland's economy and society.
A new phase in the development of corruption in Iceland (post-clientelism) began in the last decades of the twentieth century, characterized by the retreat of clientelism and replacement by a more pluralist system of professionalized politics and administration (Kristinsson 2001). This development, however, lies outside the scope of the present paper.

Conclusion
The Icelandic experience of corruption is relevant to corruption research in at least two ways. First, it sheds light on the difference between socio-cultural perspectives on corruption versus institutional ones. While Iceland shared essential socio-cultural characteristics with other Nordic countries that allegedly contributed to their low levels of corruption, it nonetheless experienced higher levels of corruption in the form clientelism in the twentieth century. Thus, socio-cultural differences are unlikely to account for differences in corruption levels. Second, weak state-building and democratization before bureaucratization are the key institutional features responsible for Icelandic exceptionalism in the Nordic context. The Icelandic bureaucracy was insufficiently developed at the start of the twentieth century to hold out against the onslaught of clientelist politicians. As the new class parties began organizing on the ground, they used the public sector to distribute favors among their clientele, building party organizations based on clientelist networks.
The essence of the story is that potential demand for corruption probably exists everywhere, irrespective of cultural traits or social structures. If the institutional features that limit the supply of corruption and enforce non-corruption are lacking, corruption is likely to prevail.
Every system of corruption, however, has some unique features and there may exists any number of syndromes (Johnston 2005) that make generalizations across contexts difficult. While the present study supports an emphasis on institutional factors to account for inter-Nordic variations, it must be added that it has, strictly speaking, not been shown that cultural accounts of corruption are in all cases inadequate.
The Icelandic case suggests that cultural features are unlikely to provide the necessary conditions for corruption since corruption developed in Iceland despite cultural similarities with other Nordic countries. Based on the present study, however, we have no way of judging if culture might be a sufficient condition, e.g. whether a highly particularistic culture might provide fertile ground for corruption, irrespective of institutional setup. Similarly, while social and cultural factors are unlikely to account for variations in corruption among the Nordic states, they might still contribute to the explanation of why the Nordic region is relatively less corrupt than most other parts of the world.

Endnotes
1 The author is grateful to participants in a workshop on Nordic political corruption organized by Lars Mjøset University of Oslo, 28-9 March 2019 where the original idea for this article emerged, participants in the corruption workshop at the NOPSA virtual congress 2021 for a lively discussion of an earlier draft, to two anonymous reviewers who made many useful suggestions and to Ásdís Björk Gunnarsdóttir for research assistance. 2 For an overview of living standards, see Ólafsson 1990. Figures on GDP presented by Guisan and Cancelo (2001) indicate that in 1964 GDP in Sweden was 173% of the OECD average, in Denmark it was 156%, Norway 132%, Iceland 129% and Finland 125%. 3 In this context it is worth recalling that even much larger administrative systems than the Icelandic one are highly network based and operate through personal contacts to an important degree (.e.g. Heclo (1977), Wildavsky (1974), Suleiman (1974). While proximity may contribute to denser networks, network theory suggests a much smaller world than large numbers might suggest. For a summary, see Barbási (2003). 4 Hálfdanarsons interpretation, it should be noted, is not uncontested. See in particular Jónsson (1995) and Karlsson (1995). 5 Hilmar Finsens was of Icelandic descent but born and raised in Denmark and did not start learning Icelandic until he was offered an administrative position in Iceland in 1865. He later became Mayor of Copenhagen and a Danish government minister -but his appointment as Governor in Iceland seems not to have reduced dissatisfaction in Iceland with the new constitutional arrangement. In Knudsen's (2000 p. 918) history of the Danish administration the appointment is described as mainly cosmetic: "For den første landshøvdings vedkommende var aendringen endnu ret kosmetisk, for han var identisk med den hidtidige stiftamtmand Hilmar Finsen, som var af Islands aet, men dårlig til det islandske sprog." STJÓRNMÁL & STJÓRNSÝSLA