Appendix I: Case Discussions

Unresponsiveness also dominated the whole period as politicization was neither successful nor fully applied. Even though politicians are notoriously known for their corrupt practices, crony capitalist behavior, and connections with criminal gangs, bribery and politicization from below are also highly frequent among low-level and high-level civil servants, locally and centrally. This not only characterized the 1990s (De Waal 2007: 8-9; Fischer 2010: 424, 426) but also the 2000s (De Waal 2007: 15; BTI 2008: 5-6; Brennan and Gassie 2009: 61-71; Mavrikos-Adamou 2014). Curiously, although purges from 1992 of top-level officials in the ministries created loyalties, most mid- and low-level bureaucrats stayed put. As a result, these former communist bureaucrats could undermine government programs continuously (Biberaj 1998: 153-154). At the same time, the tsunamis of administrative purging have created a rational for each servant to ignore the party line since they calculate being fired after the next election anyway (Brennan and Gassie 2009: 75).

ARGENTINA (1983-)

The administration was ineffective throughout the period. Territorial penetration continued despite a legacy of malfunctioning federalism and a sharply decreasing presence of administrative functions and infrastructures approaching the peripheries (see BTI 2003: 4; BTI 2006: 4). The first comprehensive legal foundation for meritocracy in Argentina was installed in 1992. It introduced a career-based, meritocratic system of recruitment and promotion throughout the federal administration (Ferraro 2004: 4). However, meritocracy was never installed for a majority of public offices. Although politicization was only formally available for around 1 % of the offices, as counted by Grindle (2010: 6), meritocracy in the 2000s lived side by side with clientelism (Grindle 2010: 15). More thorough investigations of the amount of meritocratic relative to politicized positions confirm that politicization and clientelism in hiring and firing dominated over meritocracy even in the 2000s (see e.g. Scherlis 2013: 63-64; Cingolani 2014: 122, 129, 133; Malbran 2015). Politicization continued to be perceived as vital for presidential control and survival in office. Virtually the entire senior bureaucracy was politicized and the share of politicized offices only declined moderately down the hierarchy of the administration. These practices were continuances of pre-1992 traditions only interrupted by Menem’s years of extreme clientelism and party patronage (Ferraro 2004: 7; Walker 2007: 103; see also Spiller and Tommasi 2003: 297). The same story could be told of the judiciary which, however, gained independence after Menem in a reform process started in early 1999 (Walker 2007: 104-106).

Responsiveness was highly problematic throughout the period on at least two fronts: in the workings of the bureaucracy and in the coordination and cooperation between the federal and regional levels of government (see e.g. Oszlak 2000; BTI 2003: 4). First, the administration was generally known for its high levels of inefficiency, rooted in a lack of executive legitimacy and exacerbated by rounds of politicized hirings and firings (Oszlak 2000; see also BTI 2012: 6). Corruption pervaded at all levels of the administration, and although no class of bureaucrats were pitted against the governments it was very hard for politicians to control the bureaucrats who each had personal interests blending with official duties and competing jurisdictions (Blake and Kohen 2010: 30-31). The Argentinian state was characterized as one of the most corrupt (Hinton 2005: 92-93). More importantly, this reflected uncontrolled corruption by officials across both neoliberal and more interventionist eras of economic policy (Manzetti 2014). It is therefore fair to speak of a general lack of responsiveness.
BANGLADESH (1991-2007; 2009-)

The administration was ineffective throughout the period. As for resource supremacy, problems did emerge from time to time in areas with heavy guerilla presence. However, the state had an extensive bureaucratic apparatus throughout the territory nurtured by the British colonialists and the years of union with Pakistan (Alam and Teicher 2012: 867; Mollah 2012).

Despite the British legacy in administration, there were problems separating the executive from the judicial in maintaining the judiciary’s independence in hiring and firing (Mollah 2012). The judiciary was heavily politicized (Fink 2010: 96). More generally, politicization of the civil service was extensive due to the lack of trust between political parties and the use of a quota system to mitigate ethnic conflicts in politics. The quota system systematically led to the selection of the lesser qualified at the expense of merit concerns (Alam and Teicher 2012: 867-868; Ferdous 2015). According to Jahan (2006: 9, 12), the quota system involved the forced appointment of 30 % freedom fighters, 10% women, 5 % tribal groups, and 10 % for certain districts in turn leaving only 45 % of the appointments to genuinely meritocratic criteria. But even the share of meritocratic appointments was hampered by corrupt and politicized practices in the management bureaus of the ministry of establishment and the Public Service Commission. These legacies of the British, legalized in the Government of India Act of 1919, the constitution of 1962, and the new examination procedures of 1973, were on balance circumvented more than upheld (Jahan 2006).

Responsiveness was undermined by nepotism and administrative corruption in a bureaucracy characterized by low performance, inefficiency, and bloated budgets (BTI 2003: 3; Khan 2003: 392). The esprit de corps was waning throughout the period alongside deep-seated factionalism in the bureaucracy due to the cadres that emerged from the quota system (Alam and Teicher 2012: 871). In general, the bureaucracy was known for a weak ability to formulate and implement political decisions (Zafarullah and Rahman 2008: 739).

BELARUS (1991-1994)

The administration was ineffective throughout the period. As in Ukraine, a relatively dense infrastructural network and administrative structures remained in place after independence in 1990 (Way 2005: 233; Fritz 2007: 11-12). In turn, the state escaped collapse and instead maintained a capacity for penetration (Bennett 2011: 11).

The Belarussian administrative tradition places it close to Russia relying on a strong legacy of politicization (Møller and Skaaning 2010: 325). Judicial independence was formally established by law in 1990, amended in 1992, but was never implemented as judicial appointments remained the de facto privilege of the President (Vashkevich 2012: 1065). At the same time, the administration went from politicized to outright sultanistic under Lukashenko. While most accounts focus on the last instance (under an autocratic regime), they indicate that his practice was a continuation or an exaggeration of an already existing system (see e.g. Eke and Kuzio 2000: 532; Reed 2004: 162).

There was, however, a significant strengthening in government control from the democratic to the Lukashenko regime. In the early years of the 1990s under democracy, the lack of incumbent control, as pointed out by Way (2005: 233, 249) extended to the civilian spheres of governance in the ministries and local administrations. Although extractive capacities and systems for public service delivery were stronger than in most of post-communist countries (Fritz 2007: 229-233), the quality of implementation was weak and based on patronage (Korosteleva, Lawson, and Marsh 2003: 7). There was no distinction between state and personal wealth. But whereas this was fiercely managed by Lukashenko at the top of a pyramid (Eke and Kuzio 2000: 531), the implementation of policies before 1994 was much more arbitrary (Fortin 2010: 670).

BENIN (1991-)

The administration was ineffective throughout the period. Historically, the state never penetrated the countryside fully and social control was weak (Levitsky and Way 2010: 292). The French built a highly centralized state and subsequent reforms of decentralization were inadequate and ultimately unsuccessful (Rasmussen 2013: 71). In the 2000s, the state was “still in the making” as Central Benin remained a vast territory with no state institutions in between the larger towns (Le Meur 2006). State infrastructure was only very basic and still extending by means of ineffective decentralization reforms (BTI 2003: 3; BTI 2006: 3; BTI 2012: 6).

Meritocracy was never secured. Judicial independence was only achieved formally (Bierschenk 2008:105-106). Politicization was highly dominating the appointment processes to the judiciary (Bierschenk 2008: 106, 113). A neopatrimonial, clientelist state continued after 1991 in which the state apparatus increased as oppositions had to be coopted and as the government had to achieve control over state resources and implementation (Rasmussen 2013: 74).

Neopatrimonialism also hampered responsiveness as civil servants had no education in the conduct of public office and no incentives to serve any defined public interest (Heilbrunn 1999; Bierschenk 2009). The legal system was highly ineffective with a heavy overload of cases (Bierschenk 2008: 105-106; BTI 2006: 3). And public services remained inefficiently delivered due to corruption (BTI 2006: 3). Heilbrunn (1999: 229-231) describes the infectious culture of corruption at all levels of the administration which continued after 1991 despite the purging of 80 % of the personnel.

BOLIVIA (1982-)

The administration was ineffective throughout the period. The practical working of the administration was not always present in rural areas but the state penetrated the territory. The traditional patronage and spoils systems furthermore continued (Malloy 1991: 46-47; Mayorga 2005: 166; Gingerich 2010: 55, 57), and while reforms to install meritocracy were initiated, they were never carried out systematically and vulnerable to further politicization. The 2000s under Morales saw a new wave of ministerial or presidential politicization (Dove 2002: 72, 74). Meritocracy, or a protected civil service, was thus never institutionalized at any substantial level.

Responsiveness was equally weak. Administrative corruption was pervasive as clientelism and imperfect politicization created weakly motivated servants and large pockets of virtually uncontrolled agencies (Mayorga 2005: 165; Gingerich 2010: 62). Performance in public service delivery was debilitating across governments in the 1980s, 1990s, and 2000s (Morales 2004: 220, 227). Clientelist networks ruined any coordinated government control of policies and implementation efforts (Malloy 1991: 46-47; de Lozada 1999: 71-73). This was also the case in the judiciary (BTI 2006: 9).
BOSNIA-HERZEGOVINA (1996-)

The administration was ineffective throughout the period. Territorial penetration was weak for the same reasons as for resource supremacy. Common institutional structures did exist with the presidency and the council of ministers as well as the central legislative and judiciary branches. However, real power and authority rested with the local level, at the cantons or Opstinas further augmented by the territorial split-up (Oluic 2007: 70).

The administrative legacy of Bosnia is set to be Eastern Oriental, the most patrimonial and corrupt in Eastern Europe (Møller and Skaaning 2010: 334; see also Toal and Dahlman 2011: 11). The judicial sphere was reformed in 2000 and has since been relatively independent of political pressures relying on an effective commission overseeing the hiring and dismissal of judges (Bieber 2006: 48). However, since 1996 political parties have often hijacked public offices and transformed them into bribe-collecting agencies (Divjak and Pugh 2008: 376), and civil servants were simply recruited and promoted on the basis of party political rather than meritocratic criteria (Verheijen and Rabrenovic 2015: 19).

Regarding responsiveness, it has been noted that the Dayton Accords’ move to privatize the economy and decentralize the administration only exacerbated corruption (see Divjak and Pugh 2008: 375). Cases of successful bribing of judges by private clients have remained pervasive (Bieber 2006: 48). Both before and after 1996, patronage systems were well-institutionalized to deliver ethnically targeted goods rather than public service and effective bureaucracy (Toal and Dahlman 2011: 235; Lyon 2006: 51; Ramet 2006: 483).

BOTSWANA (1966-)

The administration was effective throughout the period. The state was unusually extensive for African standards and had a strong presence throughout the territory in absolute terms – not least by the successful incorporation of local Tswana chiefs (BTI 2003: 3; Robinson and Parsons 2006). It was the continuation of governance structures from pre-colonial times making use of Tswana kingdom institutions and cooperatives with local chiefs that forged the backbone of an effective administration at independence and beyond (Henderson 1988: 220-227). The chief system was later abandoned partly but without significant problems in central management (Lekorwe 1998: 182).

The literature agrees that Botswana was an exception in Africa in terms of administrative effectiveness. It developed a strong and autonomous bureaucracy from independence that avoided any politicization being subject of a stable, peaceful multiparty system (Somolekae 1993: 113; Carroll and Carroll 1997: 473). The country without problems adopted the British civil service system and the Public Service Commission empowering civil servants against the political class – not least led by the British-educated Seretse Khama (Picard 1987: 71, 87; Henderson 1988: 229). Similarly, the judiciary was endowed with a high degree of independence and access to political firings was only given in case of gross misconduct committed by a given judge (Otlhogile 1998: 219; BTI 2012: 5).

Responsiveness was secured early on by the strength of Khama and his party, BNP, as well as the cooptation of the powerful local chiefs (Henderson 1988: 228; Somolekae and Lekorwe 1998: 190). Local-state ties were eventually successfully reformed to the benefit of the central state (Lekorwe 1998: 182). Still, from colonialism a high level civil service was maintained which avoided localization, provided high compensations, and kept well-defined accountability authority lines (Good 1994: 499). The ministries were hierarchically organized with permanent secretaries and an administrative head below effectively controlling his team of servants and delegating authority locally (Mfundisi 1998: 164). Despite signs of corruption and sluggishness in the 2000s (BTI 2008: 5), the country’s administration remained firmly effective and a model for other developing countries with committed civil servants working in peaceful relation with the political parties (see e.g. Somolekae 1993: 113; Sharma 1996: 202-206).

BRAZIL (1985-)

The administration was ineffective throughout the period. As for resource supremacy of the military and police, territorial penetration of the state administration was in place but limited in effectiveness in some areas of the Amazon forest (see e.g. BTI 2003: 4). Regarding both meritocracy and responsiveness, one must take stock of the label of ‘pockets of efficiency’ popularized by Evans (1995) in an analysis of Brazil’s bureaucracy. In terms of meritocracy, it was formally installed in constitutions for decades before the advent of democracy. This gave way to practical protection of judges in the judiciary, except at the top levels, as confirmed in the 1986-1988 democratic transition period (Rosenn 1998: 20; Santiso 2003: 164). Overall, Brazil did rank first in a Latin American list of degree of meritocracy (Cingolani 2014: 121-122, 134). Yet, this must not take attention away from the fact that meritocracy was never firmly institutionalized beyond a few, albeit core, ministries and agencies such as the finance ministry, the foreign ministry, the Central Bank, Petrobas, and the National Development Bank (Bersch, Praca, and Taylor 2013: 13; Cingolani 2014: 134). Evans himself also maintains that the ‘islands of excellence’ were surrounded by a sea of clientelism by agencies dependent on presidential protection for survival, the department of public administration declining in relevance, and generally political affiliation followed most appointments to the public sector amounting to only a very rough approximation of the Weberian ideal (Evans 1995: 61, 64). Most other country studies of Brazil reach the same conclusion (see e.g. Weyland 1998; Fernandes 1999). Reforms to improve meritocracy in the late 1990s ended up addressing personnel management more than specific recruitment rules (Fernandes 1999: 10).

Responsiveness was weak at an overall level for much the same reasons. Outside the pockets of efficiency, agencies were characterized by widespread corruption and capture by organized crime. Most seriously, coordination within the complex legal system of federalism was very weak and lawlessness was the norm in many peripheral regions in courts as well as civil services (Evans 1995: 64; BTI 2003: 4; Hinton 2005: 93).
BULGARIA (1990-)

The administration was ineffective throughout the period. As noted by Ellison (2007: 227), the socialist state in Bulgaria did not collapse in 1989 even though the government did. Throughout the post-communist period, the state thus had an extensive presence across the territory (BTI 2003: 4). The Bulgarian bureaucracy was a so-called patrimonial-communist one, a mixture of pre-communist patrimonialism and communist politicization (Møller and Skaaning 2010: 325). In the first half of the 1990s, no substantial reforms of the state administration were made. A law in 1994 indeed secured judicial independence as witnessed by numerous cases of judicial disobedience to attempts at politicization from above (Schönfelder 2005: 61-63). But the rest of the bureaucracy was characterized by frequent disposal of servants deemed ‘politically unreliable’. The pattern of patronage and politicization was altered in 1997 with the first basic civil service act. Amendments in 1998 and 1999 served to clarify further that political and administrative positions should in principle be separated (Verheijen 1999: 92, 106; Noutcheva and Bechev 2008: 131). In fact, the inauguration of Ivan Kostov’s government in 1997 has been called the beginning of state-building in Bulgaria – an indication of forthcoming improvements but also of the amount of changes needed (Bozhidarova, Kolcheva, and Velinova 1999: 1-7; Ganev 2006: 78). Tellingly, the typical assessment of the period following 1997, at least by the time of writing in 2006, is one of broken promises and inefficient implementation of meritocracy despite the creation of the ministry of state administration and administrative reform in 2003 (see e.g. Verheijen 1999: 92; Ganev 2006: 99).

Further legislative precision and instruments to realize meritocracy were adopted in 2006. According to EU reports (e.g. Shivergueva 2011: 114-116), the separation of administrative and political positions was basically successfully implemented after these amendments. The reports do not specify the ministerial variation behind this finding. In other studies, meritocracy was observed only in ‘islands of excellence’ pertaining to the ministries in direct contact with the EU such as those managing foreign relations, agriculture, and finance (Ganev 2006: 79; Andreev 2009: 388). Being a borderline case with some ambiguity regarding meritocracy levels in individual ministries, I nevertheless code meritocracy as present from 2007 relying on the EU reports and that the aforementioned ministries cover the most salient ones.

Corruption levels were some of the highest in the post-communist space (Møller and Skaaning 2010: 341). As Verheijen (1999: 92) notes, the administration in the 1990s was in crisis, implementation in a parlous situation (see also Bozhidarova, Kolcheva, and Velinova 1999: 2; Barnes 2007: 74). Judges and ministerial civil servants, being paid poorly and held at low esteem, were frequently bribed or conflicting with their superiors and administrative collaborators such as prosecutors in the judiciary over the ends and means of policies (Schönfelder 2005: 65, 72). Despite notable improvements after 1997 (see e.g. Ganev 2006: 80), legislation, strengthening of ministerial departments, and EU accession as a basic incentive for governance improvements did not change the main problems of ineffective implementation: Even after EU accession, the state was continuously captured by private actors hampering public service delivery and street-level corruption rose (Schönfelder 2005: 77-78; BTI 2006: 4-5; Noutcheva and Bechev 2008: 138; Spendzharova and Vachudova 2012: 40, 44) – something the EU itself admitted (Shivergueva 2011: 116).

BURUNDI (2005-)

The administration was ineffective throughout the period. Historically, only rudimentary but still basic administrative structures existed. However, the problem of the 2000s remained the presence of FNL strongholds which prevented the state from reaching many regions in any meaningful way (BTI 2006: 4; BTI 2008: 7-8). The FNL strongholds were basically closed by early 2009 with the demobilization and integration of their fighters into the national army (BTI 2012: 9).

Whereas territorial penetration was obtained from 2009, meritocracy and responsiveness suffered under the much more ingrained neopatrimonialism of the state bureaucracy. This not only contributed to the politicization and discrimination of ethnic identities between Hutus and Tutsis. It also politicized the appointment of servants and biased implementation as well as making implementation imprecise when political orders came from one ethnic actor to another ethnic actor (Daley 2006: 659). The peace agreement caused a system of shared powers between the ethnic groups with reserved domains of governance for the two groups but it made centralized government implementation ineffective and did not alter the neopatrimonial character of the system (Vandeginste 2009; Uvin and Bayer 2013: 263).
CENTRAL AFRICAN REPUBLIC (1993-2003)

The administration was ineffective throughout the period. The state was completely absent in many areas in terms of delivery of basic services and judicial authority (Bierschenk and de Sardan 1997). After the state collapse in 1965, civil servants were heavily concentrated in the capital (Mehler 2005: 128).

Patassé continued Bokassa’s norm of completely occupying and absorbing the state which became a model of a predatory one (Saba 2005: 183). Relatives and friends were favored in getting state jobs whereas a general bias towards Patassé’s own ethnicity was introduced (Mehler 2005: 126; Rupiya 2015: 126).

Similarly, as a trait that followed the neopatrimonial in almost mechanical fashion, responsiveness was hampered as the civil service became highly dysfunctional and paralyzed (Rupiya 2015: 121), filled with a complex web of contradictory loyalties and a motivation of the civil servants that suffered under low salaries (Mehler 2005: 128).

CHILE (1990-)

The administration was ineffective throughout the period. Territorial penetration continued through the Pinochet dictatorship and beyond (see BTI 2003: 5). Meritocracy was relatively strong in Latin American comparisons (BTI 2010: 7). For instance, the judiciary historically and from 1990 was a strongly bureaucratic organization with autonomy in personnel matters (Vaughn 1992: 593). However, there was no centralized and unified system of civil service recruitment (Marcel 1999: 304). The fact remained that even after the creation of a civil service office in 2000, many office positions were presidential appointments. 60 % of all appointments were temporary contractual and the high level bureaucracy was purged with every new government (Animat et al. 2006: 21; Hira 2014: 281).

The administration was initially highly unresponsive to the new democratic regime. This was es pecially the case for the judiciary in the 1990s as it not only resisted the democratic change that was ongoing but also the concrete policies on human rights. Its alliance with the military continued until the late 1990s. Besides this basic deficiency, the administration is usually celebrated for its efficiency in tax administration, delivery of public services including health and infrastructure, and low levels of corruption (Marcel 1999: 275-277; Bergman 2003: 595; Dammert 2009: 151; BTI 2010: 7) reflected in high stability in policies and outcomes (Animat et al. 2006: 13). So, the decisive change came around 2000 as evidenced by Pion-Berlin’s (2010: 535) account of the changing attitudes among judges with the arrest and trial of Pinochet in 1998-1999 and Laguna’s (2013: 88-89) analysis of President Frei’s successful introduction of performance management from 1994 to 2000. Therefore, I code responsiveness as established from 2000.