Substantial research has interrogated the role that quasi-democratic institutions play in elongating authoritarian rule (Gandhi Reference Gandhi2008; Svolik Reference Svolik2012). Arguably, authoritarian regimes use their ruling parties and legislatures to mitigate horizontal and vertical challenges to their survival. However, in comparative perspective, we know much less about the individuals who populate these institutions or the pathways that take them to power. Where do authoritarian regimes recruit their public officials? How does the process of recruitment reflect upon the institutional capacities of regimes to shape candidacy pools, as well as evolving contexts of authoritarian rule? Importantly, can regimes recruit collaborators who are loyal, yet simultaneously possess key competencies that regimes value? Better understanding who operates within authoritarian institutions helps us evaluate how institutions serve autocrats.
This article focuses on the legislature, which has received considerable attention as an authoritarian institution but less so in terms of legislative selection (Gandhi Reference Gandhi2008; Jensen et al. Reference Jensen, Malesky and Weymouth2014; Wright Reference Wright2008). The article highlights key challenges in legislative recruitment. Authoritarian regimes are unevenly equipped with vetting mechanisms, most notably in terms of the degree to which established parties exist that can cultivate candidates from a young age (Weghorst Reference WeghorstForthcoming). Autocrats also must adapt recruitment to evolving circumstances. Most fundamentally, the context of multiparty competition, which has come to define modern autocracy, encourages candidates who can bear certain financial obligations (Lust-Okar Reference Lust-Okar and Lindberg2009). However, the very skills that autocrats might require from their legislators can become sources of threat. Specifically, the incorporation of wealthier individuals into legislative bodies can have unintended consequences that undermine autocracy (Collord Reference Collord and Cheeseman2018).
These challenges constrain regimes differently across contexts and time. Therefore, this article focuses on evolving patterns of legislative recruitment within a single case – Cameroon. Cameroon allows for temporal comparison across two authoritarian periods. Between 1966 and 1992 Cameroon was a single-party regime under the Cameroonian National Union (CNU) and its successor, the Cameroon People's Democratic Movement (CPDM). In 1992 the regime introduced elections, but the CPDM maintained its dominance as an electoral authoritarian regime. Across these two periods there are changes in who is recruited and for what purpose, with a greater emphasis now placed on elites who can fund campaigns and supply constituency service. By the same token, these changes have proven challenging and compelled the regime to intervene in candidate selection processes. Notably, Cameroon is a case where the ruling party has historically lacked grassroots structures that could nurture legislative candidates.
The Cameroonian case also points to a potential solution to mitigate the challenges of legislative recruitment – the civil service. Civil servants, especially those in administrative roles, provide key advantages. They are highly educated and enter structured career paths that allow autocrats to evaluate them from a young age, much as an institutionalized party does. Moreover, many civil service positions offer opportunities for resource accumulation and to develop connections with distributive networks. Consequently, civil servants can offer a counterbalance to the influence of business, which arguably becomes more important under multipartyism. Indeed, since multipartyism Cameroon has retained a comparatively higher proportion of civil servants in its legislature than other African countries.
The article makes these points by accessing an original data set of over 900 legislators serving between 1973 and 2019. Data collection in autocracies is notoriously opaque (Art Reference Art2012).Footnote 1 The data were obtained from a mixture of CV records and internal biographies gathered across a three-year period of relationship building. The data do not capture a legislator's entire career but do map out major junctions with near completion. The data demonstrate significant change in the occupational composition of the legislature following multipartyism. Moreover, the data allow for exploratory multivariate regression analysis to assess whether civil servants are advantaged over other occupations, and whether this effect is more pronounced under multipartyism. The data are also used to plot out specific civil service career paths to show how both the tenure and type of position influence a legislator's legislative longevity. Combined with case-level information, this data set provides evidence for how legislative recruitment evolves across a regime's lifespan, and how institutions like the civil service possibly serve a role.
Substantively, the article fills a gap in the study of authoritarian institutions and contributes to emerging scholarship on legislative selection outside of established democracies. Previous work has, in part, depended on large-n cross-national comparisons that demonstrate the average causal effect of a legislature on regime survival (Pepinsky Reference Pepinsky2014). This article adds an actor-based perspective that highlights how legislators are cultivated through specific and imperfect processes. The article further emphasizes the role of the civil service, which has not been extensively examined as a source of legislative recruitment. Finally, the article provides one of the only longitudinal studies of legislators in an authoritarian setting. The challenging data-collection environment means that some causal propositions retain a more tentative quality. Nonetheless, such data are the only way to better assess difficult-to-observe processes within autocracies.
The challenges of recruitment in authoritarian regimes
A key perspective in the study of authoritarian politics is that dictators require the cooperation of vital societal figures, who are co-opted through promises of material rewards (De Mesquita et al. Reference De Mesquita, Smith, Siverson and Morrow2005). However, there is no sure way to enforce this bargain and prevent supposed collaborators from becoming future threats (Wintrobe Reference Wintrobe1998). Consequently, dictators engage in a variety of strategies to hedge against the risk of power-sharing. For instance, periodic purges or reshuffling can prevent any single elite from accumulating too much influence (Woldense Reference Woldense2018). The institutionalist literature argues that by empowering legislatures, more credible parameters of power-sharing are established. As Carles Boix and Milan Svolik write, ‘regular interaction in high-level deliberative and decision-making bodies … reduces asymmetries of information between the dictator and his allies’ and precludes conflict (Reference Boix and Svolik2013: 348).
Absent is more information on how collaborators are recruited in the first place. Presumably, a regime would prefer elites who have undergone a process that at a minimum reduces threat and at best produces loyalists. Historically, the presence of an established ruling party that can provide partisan avenues of recruitment has helped with this vetting. For instance, Mexico's Institutional Revolutionary Party (PRI) could rely on its close engagement with grassroots supporters to identify trustworthy candidates (Magaloni Reference Magaloni2006). In Tanzania, Keith Weghorst has shown how the ruling party uses records of partisan activism as a barometer along which to evaluate potential legislators (Weghorst Reference WeghorstForthcoming). Similarly, the Chinese Communist Party uses institutionalized recruitment mechanisms to limit selection to supporters who are clearly ‘red and expert’ (Liu Reference Liu2018).
The problem is that not all autocrats are equipped with these vetting mechanisms. This constraint is particularly evident in Africa, where frail states and societal divisions often meant that ruling parties became amalgamations of regionally based entities (Arriola Reference Arriola2009). Without clear partisan recruitment, personal reputation or informal networking through local power brokers was used instead. This is a second-best strategy since it is not always stable and risks considerable principal–agent issues. Candidates chosen based on their local connections might not serve regime ends, but rather the private motivations of their brokers. As Alex Kroeger (Reference Kroeger2020) shows, personalized African regimes are more likely to use cabinet shifts to reduce risk.
Legislative recruitment is also influenced by the skillset an autocrat needs from its collaborators. Elites might be expected to marshal basic competencies in terms of legislative process and knowledge of government, especially as legislatures become more professionalized. Ruling parties might enforce these criteria during the selection process by requiring minimal levels of education (Liu Reference Liu2018). By contrast, work on African legislatures suggests historically low levels of professionalization and a premium placed on constituency service (Mattes and Mozaffar Reference Mattes and Mozaffar2016). Legislative candidates are expected to shoulder the burden of campaigning, which demands extensive spending during primaries and general elections (Ichino and Nathan Reference Ichino and Nathan2012). Once in office, legislators are expected to deliver local collective goods such as school and road construction. Consequently, regimes might need the cooperation of wealthier individuals to sustain these financial requirements (Koter Reference Koter2017; Pinkston Reference Pinkston2016).
Finding capable legislators poses a number of challenges. First, the available talent pool might be narrow. This is again true in Africa, where professional opportunities following independence were limited to a narrow elite employed as educators or petty government officials (Hornsby Reference Hornsby1989). Second, the recruitment of wealthy individuals, primarily from the business sector, is risky. Business elites might simply be the highest bidders for their nomination, which can strain intra-elite relationships. Likewise, their legislative service could be oriented towards generating business returns rather than constituency service (Szakonyi Reference Szakonyi2018; Treux Reference Treux2014). Studies from Africa have demonstrated the importance of business in funding opposition parties (Arriola Reference Arriola2012). But, even when business is co-opted a risk remains since their relative autonomy from the state can lead to greater assertiveness (Collord Reference Collord and Cheeseman2018).
These processes are dynamic. Crucially, the shift to multipartyism creates new challenges for legislative recruitment. Competitive elections, even unfair ones, are associated with more pressure for constituency representation, as well as much larger candidate pools (Lust-Okar Reference Lust-Okar and Lindberg2009). In many cases, regimes liberalize candidate selection processes and allow a more diverse range of elites to enter the legislature. Opening candidate selection might provide benefits in terms of election results, but it is also problematic since it limits a regime's ability to control the legislative slate. This challenge is particularly acute if ruling parties were not previously used to cultivate candidates through partisan careers. Reinvigorating ruling parties to enforce stronger discipline and evaluate candidates is often costly and not a viable option.
The civil service as a source of legislative recruitment
For the purposes of this discussion, the term civil service refers to the set of individuals who enter structured career paths and hold administrative authority within the national or territorial government. While teachers and medical professionals are often employed by the state, they do not necessarily wield such authority or have access to the state. Likewise, not all who hold administrative authority obtained that power by moving up a structured career path. Instead, there are several sections of government that are managed by political appointees who enter the civil service following extensive professional or political careers. In this sense, the term civil service is nearly interchangeable with careers in civil administration.
The civil service has historically been a source of authoritarian control – particularly in Africa – but not always a pipeline into the legislature. Larry Diamond (Reference Diamond1987: 576) refers to the civil service in Africa as facilitating a ‘process of rapid class formation’, while Jean-Pierre Olivier de Sardan (Reference Olivier de Sardan1999) highlights its role in elite rent-seeking. However, only in certain cases are civil servants recruited into legislative positions. For instance, in Ghana most early legislators were traders (Kilson Reference Kilson1970: 77). Likewise, in Kenya just 8% of the 1963 legislature were civil administrators (Hornsby Reference Hornsby1989). Comparatively, the early legislatures in francophone countries had more civil servants (~33%), which is expected, given the legacy of direct colonial rule (Le Vine Reference Le Vine1968). But the available data suggest significant variation. According to the Inter-Parliamentary Union (IPU), the proportion of civil servants in contemporary francophone legislatures ranges from 0% in the Republic of Congo to 63% in Djibouti (see Online Appendix E).Footnote 2
There are a number of reasons to think that the civil service might become a useful pipeline to the legislature in an authoritarian regime. First, civil servants are often high-quality candidates (Branedle and Stutzer Reference Branedle and Stutzer2016). Recruits into the civil service usually must satisfy basic educational requirements, take entry exams or provide personal references (Liu Reference Liu2018). Certain positions require recruits to possess specific technical skills and advanced degrees, or to demonstrate minimal levels of work experience. But professional competency is more likely a desired trait in regimes where the legislature has more autonomy (Gandhi Reference Gandhi2008). Relatedly, most African legislatures are characterized by low levels of professionalization and high levels of executive dominance (Barkan Reference Barkan2009). Indeed, the IPU data suggest that in Africa there is very little representation from skilled professionals such as lawyers.
Second, the civil service can mimic the vetting processes that would have been provided by an established party. As with partisan careers, civil service careers begin at a relatively young age and are structured and hierarchical. Promotion depends on meeting observable benchmarks and vesting time. Civil servants in francophone Africa are promoted within specific categories of civil servant (e.g. from second to first degree), but also across categories of civil service (e.g. from Category B to Category A). More senior positions require appointment by a figure in the executive. This provides frequent opportunities for regimes to evaluate worthy candidates. Indeed, in Cameroon many senior civil servants simultaneously hold senior party positions, which gives them dual gatekeeping authority (Cantens Reference Cantens2010: 5). Importantly, while civil servants in democratic settings are idealized as isolated from political influence (Branedle and Stutzer Reference Branedle and Stutzer2016), in authoritarian regimes advancement in the civil service can be conditioned on political quiescence, if not outright displays of loyalty such as campaigning for specific candidates.
Third, civil service careers prepare candidates for the unique contexts of multipartyism in developing countries. Civil servants have access to material resources and at times even control over distributive means. This helps potential candidates build wealth and personal networks that can be used to secure their nomination. Once in office, legislators with administrative backgrounds might be better networked into the state bureaucracy and can therefore more easily direct the state budget back to their constituencies. Indeed, in many authoritarian legislatures, including Cameroon's, legislators have very little direct budgetary control but instead must lobby individual ministries and utilize personal networks to secure funds. Importantly, as pipelines into the legislature diversify, former civil servants can compete against business with the added benefit of emerging from a process that gives them some insulation from perceptions of threat.
This does not mean that the civil service is the only pathway into the legislature, or a panacea against the risks of power-sharing. It is in an autocrat's self-interest to bring business elites into the fold, especially since they can fund opposition parties. Autocrats can further hedge against that risk by maintaining some control over the financial sector (Arriola Reference Arriola2012). Likewise, empowered civil servants can also become a source of threat, and regimes might decide to purge them occasionally. However, this does not contradict the possibility that regimes might have an emerging preference for former civil servants as a counterbalance to the inclusion of business in elected office.
The challenges of legislative recruitment in Cameroon
This section demonstrates how Cameroon corresponds with the above-mentioned challenges of legislative recruitment, and why the civil service has possibly emerged as an important recruiting institution. First, the ruling party lacks a grassroots presence to cultivate candidates, but it has historically banked on the civil service to recruit the political elite. Second, since the multiparty era there has been more pressure to nominate candidates who can bear the financial burden of elected office and serve constituents. Third, multipartyism coincides with the liberalization of candidate selection, but also significant problems that led to regime intervention.
Cameroon's ruling party has never developed true grassroots institutions. This is primarily due to the party's amalgamated origins. After independence, first president Ahmadou Ahidjo was granted extensive emergency powers that allowed him to dictate the foundational constitution and coerce rival parties into a ‘Grand Coalition’. By 1972, even the federal structure of Cameroon was abolished, and all parties absorbed into the CNU. The CNU retained a national secretariat but little infrastructure below the section level, which corresponds with a district or department (DeLancey Reference DeLancey1987). This largely continued under Ahidjo's successor Paul Biya, even after he rebranded the CNU as the CPDM. The CPDM was created in 1985 primarily to purge the party of Ahidjo-era loyalists, as evident in the stunning 83% turnover between 1983 and 1988 (Takougang Reference Takougang2004). It was not until 2007 that party revitalization efforts restarted, when the CPDM held its first internal election in nearly a decade. It is still unclear what changes have actually occurred on the ground.
By contrast, the civil service has been an important tool of regime maintenance. This is largely due to the provisions of the first constitution, which gave Ahidjo near-exclusive powers of appointment. Figures such as governors, district officers and mayors became extensions of presidential power. Ahidjo also created the École Nationale de l'Administration et de la Magistrature (ENAM) as an elite institution to prepare students for careers in the civil service. By the early 1980s ENAM was a key means of ‘socialization into the bureaucratic-administrative class’ (Eyoh Reference Eyoh1998: 254). Importantly, following a brief period of retrenchment, public sector reforms stalled, and a higher percentage of GDP has been dedicated to public service wages annually. Some have seen in the Biya era an explicit strategy of authoritarian retrenchment by forcing legislative candidates into the civil service pipeline and curtailing independent political careers (Konings Reference Konings2010).
Across Cameroon's authoritarian tenure, the skillset required from legislative candidates has changed. Ahidjo's legislature, which in 1973 consisted of 120 members, was not a deliberative body. Ahidjo could dissolve or extend parliament and introduce bills directly into plenary without committee mark-up. Likewise, the legislature granted the president wide authority that elevated presidential decrees to the status of law (DeLancey Reference DeLancey1987: 12–13). Rather, legislators were recruited with an eye towards co-opting political opponents and representatives of the major regional-ethnic groups. This was a useful strategy since at the time the range of elites was estimated at just 1,000 (Konings Reference Konings, Jiberto and Mommen1996). Many members of Ahidjo's legislatures were once leaders of rival political parties who built their political capital in the colonial administration. Notably, legislators were selected in a nationwide constituency that limited their relationship with local supporters.
Some of this logic carried through under Biya, but multipartyism elevated the role of constituency service. By 1988, the legislature was expanded to 180 members elected in a mixture of single- and multimember districts. The expansion of single-member districts in particular has been useful in enhancing the relationship between legislator and voter. In addition, legislators were offered new fringe benefits to assist with constituency travel and fund local development projects. Much of a legislator's daily activity evolved to include bargaining with government agencies over project funding and demonstrating constituency-level improvements (Ntamack Reference Ntamack2011). In contrast with the single-party era, the legislature is no longer simply a rubber-stamp institution, but it is also not a true deliberative body. The legislature has little technical staff and passes very few bills, the vast majority of which originate in the executive (Ntamack Reference Ntamack2011).
Crucially, the shift to elections corresponds with significant change in how elites access the legislature and who is represented. Reflective of the narrow elite at the time and the focus on regional-ethnic balancing, candidate selection between 1973 and 1983 was centralized within the CNU's small Political Bureau (DeLancey Reference DeLancey1987). The civil administration and education sector – the two major sectors of employment – remained the key occupational pathways to the legislature. Yet, by 1988 Biya had liberalized candidate selection and allowed local-level primaries with delegate voting. Restrictions were minimal, and any literate 23-year-old who paid party dues could run. The consequence was an explosion in candidacy, ranging from 1,200 to 3,000 people per election. At the same time, more relaxed financial restrictions and some economic liberalization led to the emergence of a considerably larger business class.
These changes proved problematic. Many business elites were initially opposition supporters and therefore not natural regime partners (Arriola Reference Arriola2012). Concurrently, CPDM militants frequently accused wealthy candidates of vote buying during primaries. The CPDM's central committee issued circulars that specified additional criteria such as ethical behaviour or demonstrable loyalty, but these were unevenly enforced. Likewise, the central committee was reluctant to veto primaries, given the risk of local backlash (Morse Reference Morse2019). By 2013, primaries were abolished and replaced with a semi-centralized system. Constituencies could submit competing lists of candidates and even run an informal primary, but the central committee vetted all files and endorsed the eventual nominees. This gave the regime more control over the legislative slate while remaining attuned to constituency demands. As noted below, this also corresponds with a substitution of business elites for former civil servants.
Occupational background and legislative selection in Cameroon
This section uses the biographical dataset to plot changes in the composition of Cameroon's legislature, and to explore whether the regime now possibly prefers candidates with civil service backgrounds. The analysis is limited to ruling party legislators, leaving a data set of 637 individuals (n = 923). The occupational classifications build on similar work, and codes where a legislator spent the bulk of their previous careers. There are very few examples of perfectly mixed careers, which are categorized under ‘Other’. For some categories, finer-tuned distinctions are made based on the degree of seniority or specialization (see Online Appendix D). With regard to civil servants, as noted previously, this category only includes individuals who served in the administrative arm of the state and emerged from structured career paths. This helps focus the analysis only on those individuals who truly emerged from the civil service.
One further caveat is that this analysis only examines legislators who sought and obtained additional terms in office. This determination was made based on nomination records and secondary sources. The bulk of ruling party legislators seek renomination, and on average just 8% retire or die. There is no significant difference across occupations, and importantly business elites seek renomination at rates similar to that of civil servants. However, with the exception of 2013, the data cannot determine whether someone lost their primary or was removed by the party. Moreover, only sitting legislators are observed and not the range of candidates within the ruling party who seek elected office. Therefore, comparisons can only be made regarding how losing incumbents stack up against those who are nominated for additional terms.
The changing composition of Cameroon's National Assembly
As summarized in Table 1, the occupational composition of Cameroon's National Assembly has undergone significant changes. Reflective of the limited elite, the bulk of legislators between 1973 and 1983 were either teachers, junior administrators or clerks. Very few were employed in business or in specialized professions such as medicine or law. Crucially, for the most part these were not elites that Ahidjo had cultivated himself, but rather those he brought into the legislature following the abolition of federalism. Over a third were active in the post-independence era and served in the colonial or territorial assemblies, and many held traditional titles. Their occupation as civil servants placed them in the political elite, but they were not forged by the civil service to become regime loyalists. Indicatively, these legislatures were comparatively stable, and nearly all members were selected for consecutive terms.
Table 1. Occupational Composition (%) and Turnover Rates (%) of Ruling Party Legislators, by Legislative Session
Note: Data are from author's biographical data set. Numbers in parentheses are the rate of renomination for each occupational sector.
Subsequent changes in the occupational composition of the legislature initially reflect the transition from Ahidjo to Biya. Between 1983 and 1988 all occupational sectors experienced declines in renomination rates, with junior administrators suffering the brunt. Just 17% of junior administrators who sought additional terms were renominated, and their total proportion dropped to a nadir of 8%. Most of these former civil servants were remnants of the Ahidjo-era elite, from which Biya sought to extricate himself. Indeed, of the 27 legislators who began their careers prior to 1972, Biya retained just three. By contrast, by 1988 the proportion of business elites in the legislature began to grow in tandem with the start of economic liberalization.
Pipelines into the legislature diversify greatly following the transition to multipartyism. Between 1992 and 2007 the proportion of businesspeople increased to a peak of 36%. Moreover, the range of business endeavours represented expanded, with more individuals who heralded from large resource extraction, commercial development and manufacturing firms. Similarly, senior professionals have also emerged as a significant occupation, but notably with very few legal professionals (see Online Appendix B). In tandem, the currency of work in comparatively less lucrative occupations such as education has declined considerably, as has the cachet associated with traditional titles.
The multiparty period initially harmed civil servants, who by 2002 constituted 21% of the legislature. However, this proportion is still high compared with other African nations (see Online Appendix E). Moreover, the proportion of civil servants rose again to 27% in 2013. This trend is driven by an increase in senior administrators and civil servants who spent considerable time in a civil service controlled by the Biya administration. Notably, the 2013 change to candidate selection has had an apparent effect. Of the 51 legislators dismissed between 2007 and 2013, 39% were businesspeople while just 16% were former civil servants; 43% of their replacements were civil servants, while just 20% were businesspeople. This implies a deliberate strategy and an emerging preference for legislators with civil service experience, likely in response to the growing influence of business.
Civil service backgrounds and legislative renomination
This section uses multivariate regression to explore how civil service backgrounds impact the odds of renomination. The expectation is that controlling for several individual- and department-level factors, civil servants are more likely to be renominated compared with other occupational backgrounds given their emergence from an established vetting process and access to resources. These effects are likely to be pronounced under multipartyism when the relative advantage of civil service vis-à-vis business can be observed and is expected given the challenge to nominate loyal candidates who can self-fund. The analysis compares career civil servants and businesspeople against all other occupational categories. A multilevel mixed-effects regression is used, with year observations nested into individual legislators, who are themselves nested into departments (Rabe-Hasketh and Skrondal Reference Rabe-Hasketh and Skrondal2012).Footnote 3 The results reported are the odds ratio of renomination given marginal change in an independent variable.
A number of individual-level measures are included. First, signifiers of legislative status such as service on the finance committee or an official legislative leadership position are included. Both factors provide legislators with advantages through the greater access to distributive resources. Moreover, allocation of these roles might account for pre-existing prestige. Second, measures of previous political careers are included by noting whether a legislator won a local election or was provided with an executive appointment. These factors could signify stronger community ties or regime connections. Third, during multiparty election years, measures were added for competition in a single-member district and in an historical regime stronghold where intra-party competition is stronger (Letsa Reference Letsa2017). Fourth, standard biographical markers account for gender, age elected, education and traditional title. Finally, the number of terms (and its square) was included, as well as a dummy variable for each election year.
Other measures account for some department-level heterogeneity. First, the natural log of population density provides an estimate of the extent to which a department is rural, and therefore potentially less economically developed. In rural areas, there might be fewer legislative candidates and higher social dependency, which would reduce turnover. While an indirect measure of development, population density is the most consistent across time. The supplementary materials include models that account for the proportion of homes with mud floors and without electrification, although this data is very limited. Second, a rough assessment of the department's ethnic heterogeneity is made. Ethnic heterogeneity is predicted to increase turnover rates given intergroup pressures to rotate access to higher office. Summary information on all variables can be found in Online Appendix A. The results are reported graphically in Figure 1 and the specific model outputs are reported in Online Appendix C.
Figure 1. Multilevel Mixed Effects Logistic Regression Comparing Single-Party to Multiparty Era
Notes: Odds ratio of legislative renomination by the CPDM is reported. Grouping variables are department (level 3), legislator (level 2) and year (level 1). All models include control for year.
The results demonstrate that under multipartyism the odds of renomination for former civil servants are significantly higher (2.1) than for other occupations. This finding holds when controlling for 2013, when the regime signalled its most explicit preference for civil servants. Primary voters arguably view civil servants as capable representatives who can effectively interact with the state and perhaps also have the regime's implicit endorsement. By comparison, businesspeople have 1.6 higher odds of renomination than other occupations, but the impact is not statistically significant. Notably, as expected, these dynamics do not hold in the single-party era. While the civil service was an important career, it did not yet have a comparative advantage over other occupations. Importantly, in the models reported in Online Appendix C the interaction of multipartyism and civil service occupation is statistically significant, indicating that the difference between the single- and multiparty eras is meaningful.
A number of other variables also exert influence. Service in the executive branch substantially increases the odds of renomination. Notably, the impact is insignificant during the single-party era, when there was more movement between the executive and legislative branches. Similarly, winning a local election for councillor or mayor also significantly increases the odds of legislative renomination under both single- and multiparty conditions. Service on the finance committee increases the odds of nomination, but only during the single-party era. This might be due to the fact that the committee had more clout then. Since 1992, the legislature has included an overlapping economic affairs committee. By contrast, leadership roles exert more influence during the multiparty period. This is likely because of the expanded fringe benefits now associated with these positions. Demographic factors besides age elected wield no influence, and department-level variables are insignificant.
The results also show that during multipartyism competition in single-member districts is associated with higher odds of renomination compared with multimember districts (2.1). This makes sense, since in single-member districts there are stronger incentives to cultivate a personal vote, while in multimember districts legislators depend more on party connections (Carey and Shugart Reference Carey and Shugart1995). This is mitigated somewhat in Cameroon by the fact that in some multimember districts, legislators are appointed to represent specific constituents. For instance, Mfoundi elects six legislators, but each oversees a section of the department. In Online Appendix J, occupation was interacted with district magnitude and the results were insignificant while the odds ratio of civil service remained significant. The conclusion is that district magnitude likely exerts a slight independent effect, and that occupational background influences the chances of renomination similarly in both single- and multimember departments.
It is important to note the limitations in the data. First, there is no available information on a legislator's actual service. Ideally data on constituency-level provision such as road construction or publicly accessible records on legislative productivity would be available. This means we cannot directly assess whether legislators are renominated because they are better performers, or whether civil service is a proxy for these qualities. Online Appendix J includes models that control for an individual's electoral performance, which is perhaps a marker of popularity but not necessarily service. Still, as noted, legislative competency is likely not a central trait required in Cameroon. Second, there are limited data on partisan careers, so we cannot say with complete certainty whether the party does not cultivate future legislators. What is available is reported in Online Appendix I and shows that CPDM subsection leaders are possibly more likely to be renominated. However, the absence of these data – especially prior to 2007 – is itself evidence of the weakness of the party. Third, we can only approximate the time served in government and the degree of seniority across the entire data set. Models reported in Online Appendix H do this and show that senior administrators are at a slight advantage, as are those who spent a larger proportion of their pre-legislative careers in government.
Unpacking civil service pathways into the legislature
The prior discussion emphasized the utility of the civil service during multiparty elections in providing candidates with financial resources and familiarity with the state bureaucracy. But not all civil service careers provide these benefits equally. Moreover, it is also possible that longer processes of nepotism are at play that determine both success in the civil service and selection to the legislature. This section proceeds under the assumption that the type and length of civil service tell us something about how civil service matters. The analysis looks at first-term legislators elected during the multiparty period, which limits the comparison to individuals on similar footing and addresses the fact that the data set is right censored. Given the smaller sample size (n = 76), what follows is primarily descriptive.Footnote 4
Career civil servants begin in either the national or territorial administration. As they progress, some remain in junior roles while others are promoted to senior positions. Both junior and senior civil administrators can spend similar lengths of time in government, and therefore can be subject to similar vetting. However, senior administrators are expected to have access to more material resources and better networking within the state bureaucracy. Moreover, several senior administrators obtain significant decision-making roles such as director general of a ministry. These roles generally require executive appointment and provide unique advantages to would-be legislators. Service in a senior or executive role can come after a lengthy career, which would indicate that legislative candidates emerge from the process of civil service itself. By contrast, if someone obtains a senior role fairly quickly, it suggests that some pre-existing factor has influenced both their accomplishment in civil service and selection for the legislature.
Two subsets of civil service are also important to consider. First, civil servants who spend their careers in territorial administration as divisional officers are perhaps advantaged because of their stronger ties to local gatekeepers. Local government officials are often tasked with implementing public policy, and therefore have significant control over distribution. Second, financial controllers are a subset of administrative positions worth noting for the simple reason that so many legislators self-report them. Financial controllers are civil servants in a treasury or customs office, which gives them ample opportunities for wealth accumulation and therefore advantages them over other civil servants. Indeed, studies of the civil service in Cameroon have termed custom officials the ‘aristocracy’ of the civil service (Cantens Reference Cantens2010).
Table 2 summarizes these career pathways. The majority of first-term former civil servants began their careers as junior officials in the national administration (74%), followed by financial controllers (18%) and territorial officers (8%). Approximately half were subsequently promoted, and of those most ended up in senior administrative or executive positions. Fewer civil servants began and ended their careers in financial control or territorial administration. In total, by the time they were elected to office, the majority of former civil servants were in junior administrative roles (32%), followed by executive roles (23%), senior administrators (14%), senior financial controllers (12%), junior financial controllers (8%), senior territorial administration (7%) and junior territorial administration (4%).
Table 2. Civil Service Career Pathways Among First-Term Legislators (1992–2019)
Figure 2 plots these first-term legislators over their type and length of service and notes their average renomination rates. First, the majority of civil servants only enter the legislature after dedicating at least 10 years, including those in executive positions. Overall, approximately 3% spent 10 years or less in government, 46% between 11 and 20 years, and 51% over 20 years. This indicates that serving in the civil service produces something of value for legislative candidacy. In terms of renomination rates, higher seniority is influential. Senior administrators with 11 to 20 years of service are nearly twice as likely to be renominated compared with junior administrators with similar tenures in government. While a much smaller subset of civil servants, territorial administrators enjoy 100% renomination rates after their first term. Finally, financial controllers are also advantaged, primarily compared with more generic junior administrators.
Figure 2. Type and Length of Government Service and Legislative Renomination.
Note: Type of government service is broken down according to length of service (1 = 0–10 years, 2 = 11–20 years, 3 = >20 years).
A number of narratives help to flesh out these career paths. For instance, Théophile Baoro exemplifies a career spent in financial control. After graduating from ENAM, Baoro rose through the ranks at the Treasury Department, before being sent to act as Power of Attorney for the Treasury in Douala, and later to act as the financial receiver for the city of Yaoundé. Baoro was elected for three terms in 1997 at the age of 40, after spending 15 years in government. Similarly, three-term legislator Abakar Mahmat spent over 20 years working as a customs official, rising from a junior executive officer to a senior inspector. On the other hand, Sali Dairou presents a career that spans the junior administration to an executive role. After graduating from ENAM, Dairou served nearly 26 years in the civil administration, eventually rising to become a director of budget and later assistant director general of the Ministry of Finance. He went on to sit in the legislature from 2002 to 2019. Likewise, Luc Koa began a career in 1986 as a basic posts and telecommunications officer and became a regional post and telecommunications inspector before serving in the legislature between 2002 and 2019.
This does not mean that senior civil servants are universally advantaged. For instance, civil servants with over 20 years of experience are less likely to be renominated than those with 10–20 years’ experience. This is congruent with evidence from Africa that suggests the legislature offers a retirement platform for certain senior civil servants (Woldense Reference Woldense2018). On the other hand, the 1997 nomination of a figure like Hamadou Sali shows that nepotism is also relevant. Sali was a first degree executive officer, a fairly junior position, and elected at the age of 32 after spending less than eight years in the civil administration. Similarly, Genesis Mbucksek began his career as a teacher, and transitioned to become an education inspector, where he spent approximately just 10 years. Mbucksek never rose above this junior rank but has been a key CPDM ally in the oppositional north-west.
Conclusion
Much less is known about recruitment in authoritarian regimes, especially across time. Yet, understanding these processes is essential if we are to understand how quasi-democratic institutions serve authoritarian ends. This article claims that legislative recruitment is challenged by a number of factors such as the availability of a vetting mechanism, the skillset required from legislators, and the possible contradictions that might arise from including specific types of candidates. To better understand these challenges requires closer interrogation of individual cases and interaction with difficult-to-obtain data on individual legislators. The case of Cameroon was used to highlight evolving processes of legislative recruitment within the context of a weaker ruling party and growing pressure for self-financing candidates who can provide constituency service. The case also highlighted the potential role that the civil service can play in certain circumstances in legislative recruitment. In Cameroon, the civil service has remained an important pathway into the legislature, and one that possibly provides unique advantages to candidates and the regime.
The study suggests a number of directions for future work. First, it is important to consider how authoritarian institutions overlap. Autocratic institutions such as legislatures can only be understood in conjunction with the institutions used to recruit – whether a ruling party, civil service or something else. Second, future work should consider mixed strategies of recruitment. There are multiple pathways into the legislature, not all equally vetted by an authoritarian regime. Finally, the article highlights the utility and difficulty of using actor-based accounts. This data collection is challenging but notably absent in autocracies. Since data collection is likely to be incomplete, there are limitations in our ability to adjudicate decisively on certain causal claims. Therefore, it is important to supplement these data with plausible theory and rich case-level information that ultimately leads to more contextual work. But, without these data we have an incomplete understanding of key issues such as the sources of authoritarian durability, the nature of representation and the potential for legislatures to engender democratic reform.
Supplementary material
To view the supplementary material for this article, please visit https://doi.org/10.1017/gov.2021.10
Acknowledgements
The author would like to thank Veronique Ntamack, Fru Norbert Suh I and Youssouf Bellamy for their invaluable research assistance. Thanks also go to Noah Nathan and Dominika Koter for their in-depth comments on conference presentations. A draft of this paper was presented at the ‘Parliaments and Democracy in Africa’ workshop held in Dakar, Senegal, 17–20 March 2019. I thank all the participants for their support and insights. Gratitude also goes to the anonymous reviewers from this journal for their generous suggestions. The research was made possible by funding from the University of Connecticut's Scholarship Facilitation Fund (SFF).
