A state’s ability to deliver public services to its citizens affects the development of individuals, communities, and economies. Government investments in physical inputs (e.g. offices, a reliable electricity supply, and computers) influence how public sectors operate, as do the financial resources available to politicians and bureaucrats for spending on public programs. However, the day-to-day actions of individual politicians and bureaucrats determine how efficiently public resources are used. Recruiting bureaucrats based on merit is widely regarded as a crucial component of effective and efficient governance.Footnote 1 Skilled bureaucrats possess the experience and knowledge necessary to engage in tasks such as planning, budgeting, procurement, and monitoring, which facilitate the smooth delivery of public services.
However, meritocratic recruitment cannot ensure high-quality governance on its own. Once they are hired, bureaucrats seek to advance in their careers. When politicians have the authority to interrupt their career progression, bureaucrats may sacrifice efficient public service delivery to satisfy the demands of their political principals. Politicians can manipulate bureaucrats’ professional progression using what I call “career control,” which includes interfering in bureaucrats’ promotions, their day-to-day work tasks (or “work portfolios”), and the geographic locations where they are posted. A central claim of this book is that politicians deploy career control tools not to improve bureaucratic performance, but to manufacture loyalty. That loyalty, in turn, facilitates the non-programmatic distribution of resources and, at times, the illicit capture of state assets. I theorize that these tactics will be especially prevalent in electorally competitive districts, where politicians are more likely to rely on non-programmatic strategies to secure votes and where the costs of campaigning are higher. The political co-optation of bureaucrats through career control ultimately produces suboptimal outcomes for voters, including low-quality public infrastructure and the inefficient allocation of public goods.
This book shines a light on the grey area in which most states arguably lie – between a completely politicized state and one governed by meritocratic principles across bureaucrats’ careers, from recruitment and promotions to postings. Bureaucratic management is a vital and understudied determinant of state performance. The power relationships between politicians and bureaucrats are essential to understanding within- and cross-country variation in governance.
Focusing on the case of Ghana, the empirical chapters assess the observable implications of the theory advanced in Chapter 2: (i) that meritocratic recruitment alone is insufficient to shield bureaucratic processes from political interference; (ii) that the use of career control tools is associated with worse governance outcomes; and (iii) that conditions fostering non-programmatic distribution and high electoral costs – such as intense local electoral competition – lead to more political interference in bureaucratic careers and, as a result, greater distributional bias and corruption. Chapter 3 demonstrates that hired bureaucrats increasingly possess high levels of education and are selected through a recruitment process that includes ostensibly merit-based procedures, such as examinations. The subsequent chapters demonstrate that despite such procedures, poor governance practices are widespread. Chapter 4 shows that just less than half of local bureaucrats (46%) agree that uncompetitive procurement processes that favor the incumbent party occur at their local government. Furthermore, and in support of the second observable implication, results from a list experiment demonstrate the link between career control and malfeasance, with 58 percent of bureaucrats expecting that whistleblowing about administrative misconduct would result in a geographic transfer.
Chapters 5 and 6 analyze the real distribution of public procurement contracts and local public goods projects within and across districts. Chapter 5 demonstrated high levels of market fragmentation and geographic concentration in the awarding of contracts: most companies won only one contract during the considered years, and over 80 percent of firms won contracts from a single local government. Overall, the data suggests that firms need strong connections to a district’s local mayor to win a contract in a particular district. Chapter 5 also showed that when uncompetitive procurement occurs, the types of firms that benefit are those either owned by a party executive or a donor to the ruling party. A survey experiment further demonstrates that party-affiliated firms can win contracts even without the requisite experience.
Working with bureaucrats to link actual infrastructure projects to communities, Chapter 6 explored the allocation of projects within districts. The analysis of real projects, combined with the results of a survey experiment with bureaucrats, provides evidence for the biased distribution of public resources: communities that voted for the ruling party in the previous election are more likely to receive new community infrastructure projects. In other words, local politicians appear to engage in the non-programmatic distribution of local public goods.
Considering variation in governance practices across districts, and the third observable implication of the theory, Chapters 5 and 6 provide evidence that the biased allocation of resource and uncompetitive awarding of contracts are especially likely to occur in districts that have (i) high levels of electoral competition and (ii) where the mayor is politically ambitious. I argue that these conditions provide local politicians with increased motivation to threaten – and use – career control tools against bureaucrats.
Chapter 7 addresses two key questions that emerge from the empirical findings. The first is the extent to which my theory applies beyond Ghana. I expect politicians in many other democracies to have incentives to utilize career control tools to coerce bureaucrats primarily because non-programmatic politics is a widespread feature of politician–voter linkages in most low- and middle-income democracies. When voters reward politicians for distributing personal and community benefits (rather than universal ones), politicians have an incentive to limit career separation to retain discretion over the distribution of public resources. Furthermore, elections are costly in many parts of the developing world, and political candidates across varied contexts are expected to self-fund their campaigns. I examine the cases of India and Indonesia, which, like Ghana, are electorally competitive, middle-income democracies. I find evidence of similar dynamics between politicians and bureaucrats in these countries as I find in Ghana, which suggests my argument has broad external validity.
The second question I address is whether, and how, the equilibrium I describe can be destabilized. I begin by examining reforms that more clearly separate the careers of politicians and bureaucrats. I then consider supply- and demand-side factors that might encourage a shift toward programmatic politics. Next, I assess whether political finance reforms could reduce politicians’ incentives and opportunities to deploy career control tools, thereby improving governance outcomes. I conclude the chapter by outlining promising directions for future research.
7.1 Shadow Cases: India and Indonesia
Where voters expect politicians to provide non-programmatic distribution (and reward them for doing so), politicians have an incentive to control bureaucrats’ behavior. Accordingly, I expect the dynamics I describe in Ghana to apply in other democratic contexts where voters reward politicians for supplying them or their communities with discretionary benefits. In Section 7.1, I present India and Indonesia as two shadow cases. In both cases, I find evidence that (i) politicians use career control tools and (ii) that they do so for similar reasons as local politicians in Ghana. The fact that these cases lie outside sub-Saharan Africa suggests that my argument has broad external validity and is not region specific.
Table 7.1 compares Ghana, India, and Indonesia in terms of population, wealth, political rights, the prevalence of the distribution of gifts during elections (an indicator of non-programmatic politics), and an indicator of whether political parties receive public funding. India is the most populous country in the world, with over 1.4 billion people, compared to 275 million in Indonesia and 33.5 million in Ghana. India’s voting population is similarly expansive: nearly a billion people were registered to vote in the country’s 2024 national election. Indonesia registered just over 204 million voters for its elections the same year.
| Country | GDP per capita | Population rights | Political | % voters who receive gifts | Public funding of parties |
|---|---|---|---|---|---|
| Ghana | $2,363 | 33,475,000 | 35 | 11 | No |
| India | $2,257 | 1,417,173,000 | 33 | 33 | No |
| Indonesia | $4,333 | 275,501,000 | 30 | 34 | Yes |
Source: GDP per capita (2021) and population statistics (2022) from the World Bank. Political rights (out of 40) (2023) from Freedom House. The share of the population that receives gifts from candidates/parties, data from Afrobarometer (Ghana), Bussell (Reference Bussell, Kapur and Vaishnav2018) (India), and Muhtadi (Reference Muhtadi2019) (Indonesia). Public funding of parties from International IDEA’s Political Finance Database (www.idea.int/data-tools/data/political-finance-database).
The World Bank classifies all three countries as middle income: Indonesia (upper middle) is the wealthiest ($4,333 per capita), and Ghana and India (both lower middle) are poorer, with similar income levels ($2,363 and $2,257 per capita, respectively). Each country ranks similarly in terms of democratic rights, and all hold regular elections that are largely considered free and fair: they have led to several peaceful turnovers of power. These electoral conditions incentivize politicians in all three countries to value the recruitment of competent bureaucrats.
Yet, non-programmatic politics is a salient feature of electoral politics in each of these countries. While self-reported data on accepting gifts from political candidates may not be completely trustworthy, estimates suggest that 11% of voters received a gift during recent elections in Ghana,Footnote 2 versus 33% in Indonesia.Footnote 3 While counterpart statistics from India are harder to find, national parliamentarians estimate that 34 percent of voters receive money or other gifts during election campaigns.Footnote 4 Similarly, a survey conducted in 2018 found that just over 40 percent agreed that candidates’ distribution of “cash, liquor, or gifts” was an important factor influencing voting decisions.Footnote 5 Ahead of the 2024 election, the country’s Electoral Commission seized assets from candidates worth more than 47 billion rupees ($550 million), including alcohol, drugs, cash, and precious metals.Footnote 6 These statistics suggest that non-programmatic distribution is a salient feature of elections in India.
In Chapter 2, I describe how pressures to deliver community or individual benefits to voters can encourage politicians to coerce bureaucrats to redirect or capture public resources. I also explain that pressures to capture public funds are exacerbated when elections are costly, especially when individual political candidates bear these costs. Table 7.1 indicates that only Indonesia provides public finance to political parties. However, these funds are relatively inconsequential for national organizations, so candidates are still under significant pressure to raise funds personally.Footnote 7
7.1.1 Meritocracy in India and Indonesia
Similar to Ghana, India and Indonesia have taken multiple steps to professionalize their public sectors. However, as in Ghana, there is also evidence that politicians use career control tools to co-opt bureaucrats, and both countries experience significant challenges to good governance.Footnote 8 In India, bureaucratic transfers have been called “the most visible and lamentable” tool that politicians use to manipulate the running of the public sector.Footnote 9 In Indonesia, politicians interfere in both bureaucrats’ promotions and transfers; politicians find it particularly beneficial to co-opt bureaucrats since the latter exercise independent control over public resources.Footnote 10 I begin Section 7.1 by providing an overview of the bureaucratic structure and hiring practices in both countries.
India’s Public Sector
India’s vast public sector employs about 30 million individuals. In 2015, public sector jobs constituted an estimated 60 percent of jobs in the formal sector.Footnote 11 Bureaucrats serve at five levels of government: central, state, district, block, and village. Most work at the state level.
The British colonial government introduced competitive exams for entrance into the Indian Civil Service. Indians were only allowed to sit the exam beginning in 1922;Footnote 12 the Congress Party maintained this structure after independence in 1947. It now recruits top-ranked civil servants through competitive exams administered by the Union Public Service Commission (UPSC). The “higher” civil service consists of the Indian Administrative Service (IAS), the Indian Foreign Service, the Indian Police Service, and the Central Services (Classes I and II).Footnote 13 There are roughly 4,800 IAS officers, who head all government departments at the central and state levels.Footnote 14
To take the entrance exam to enter the higher civil service, most candidates must be aged 21–30 years and have a degree from a list of specified universities or a technical or professional degree approved by the government.Footnote 15 Applicants from reserved groups (Scheduled Castes, Scheduled Tribes, and Other Backward Classes) are eligible up to thirty-five years of age. The selection process is increasingly competitive: there were 436 applicants per place in 1992Footnote 16 and 4,000 applicants per position in 2015.Footnote 17 The number of times most candidates can sit the entrance exam is also capped at six; successful recruits average four attempts.Footnote 18
While the top 100 or so qualify for the IAS, the remaining candidates are eligible for entry into the “All India” and Central Civil Services.Footnote 19 Once IAS officers are recruited, they become part of the state cadres. The public service commissions of individual states recruit into state-level bureaucracies following processes that are similar to the UPSC.Footnote 20
According to data from the QoG Institute, referred to in Chapter 1, experts agree that formal examination results guide the selection process; score average 6 on a 7-point scale. IAS officers’ careers follow a clear system of promotion.Footnote 21 However, as discussed below, postings to particular locations are often subject to political interference.
Indonesia’s Public Sector
Indonesia proclaimed independence from the Netherlands in 1945. While the colonial government operated with relatively few bureaucrats, the number of positions expanded rapidly after independence. The bureaucracy was highly politicized during Suharto’s “New Order” regime: civil servants were forced to be members of the ruling Golkar Party and tasked with persuading ordinary citizens to support the party.Footnote 22 The onset of democracy in 1998 led to significant attempts to depoliticize and professionalize the public sector. Early legislation (Law No. 43/1999) codified a merit-based approach to personnel management, and later legislation (Law No. 5/2014) established support for performance-based evaluations and incentives for civil servants.Footnote 23
Indonesia’s public sector currently employs about 4.5 million individuals.Footnote 24 Fewer than a quarter of these bureaucrats work in the central government (22%); the rest work across the country’s 34 provinces and 508 local governments.Footnote 25
While the recruitment system is not perfectly competitive,Footnote 26 important steps have been taken to limit political interference. For top appointments, the 2014 Civil Service Reform Law (Law No. 5/2014) created a Civil Service Commission with government, academic, and civil society members.Footnote 27 This commission is tasked with recruiting elite bureaucrats, including from outside the public sector.Footnote 28
The Ministry of Administrative and Bureaucratic Reform centrally manages the country’s permanent civil servants. It determines the number of public sector positions at all levels.Footnote 29 The Badan Kepegawaian Negara, the civil service agency, is responsible for recruitment; it administers two computer-assisted tests. This centralized control over the number of positions and the recruitment of permanent civil servants decreases the extent to which politicians can trade permanent positions for patronage.Footnote 30 As a result, politicians sometimes attempt to use temporary jobs as patronage.Footnote 31 As I discuss below, political interference often occurs through actions that politicians can take after bureaucrats have been hired, especially transfers and promotions.Footnote 32
The civil service is an attractive career path for college graduates, and recruitment is highly competitive: over 3 million individuals apply each year.Footnote 33 Applicants must apply to a specific advertised vacancy. After a pre-screening process, selected applicants take a basic competence exam. Applicants know their score for this exam immediately: to address allegations of post-examination score fixing, scores are projected outside the exam room in real time.Footnote 34 The top three performers on this test sit a second (typically computer-assisted) test that assesses specialist knowledge related to the position. The specialist test carries a weight of 60%, while the basic test is weighted at 40%.Footnote 35 The top-scoring candidate is selected for the position.
As in India, QoG Institute data indicates that exam performance influences final recruitment. The average score among experts asked whether exam results guide hiring was 6.24 out of 7. Once recruited, public sector positions are structured hierarchically into multiple echelons. Advancing through these echelons requires promotion, which is theoretically based on merit.
7.1.2 Politicians’ Use of Career Control Tools in India and Indonesia
While permanent bureaucrats are recruited through competitive and meritocratic channels in India and Indonesia, politicians in both countries have been found to employ career control tools to extract bureaucratic loyalty. In India, they have used discretionary geographic transfers for at least forty years.Footnote 36 Surveys of Indian politicians at various levels of government reveal that most politicians (62% of national legislators and 43% of state legislators) believe they can transfer bureaucrats,Footnote 37 particularly those at lower levels. Over 90% of national legislators who said they had the power to transfer bureaucrats asserted that they could transfer state-level bureaucrats working in their constituency, compared to just over 60% who believed they could transfer IAS officers in their constituency.Footnote 38
Data on bureaucrats’ placements over time indicate significant rotation, which suggests that transfers are widely used – and thus a credible threat to bureaucrats. An empirical analysis of transfers among IAS officers establishes that in a given year, officers have a 53 percent chance of being transferred; they are shuffled every sixteen months, on average.Footnote 39 There is evidence that such transfers are often politically motivated. For instance, IAS officers are 10 percent more likely to be transferred when a state elects a new chief minister.Footnote 40 Furthermore, transfers are predicted not only by bureaucratic competence but also bureaucrats’ ties to politicians. Bureaucrats who are co-ethnics with ministers are more likely to be transferred into desirable positions.Footnote 41 A survey of bureaucrats reports that lower-level (district) officers were transferred an average of 2.6 times over five years, while block officers reported 2.7 transfers during this period.Footnote 42
As in Ghana and India, meritocracy in Indonesia co-exists with informal processes that undermine bureaucratic professionalism. Local politicians such as mayors exercise particularly significant discretion over bureaucrats’ careers – mainly by interfering in promotions and transfers. Politicians often use promotions as “carrots” to reward bureaucrats for their loyalty; they are increasingly common immediately before and after elections.Footnote 43 There is also qualitative evidence that bureaucrats sometimes pay to be promoted. Accordingly, politicians sometimes attempt to curry favor with them by reducing or waiving these costs.Footnote 44
The transfer of bureaucrats is referred to locally as rolling pejabat (the rolling of officials); district heads have significant control over their movement. Transfers are used as both “carrots” and “sticks” to encourage compliance. For example, politicians employ transfers to reward (or punish) bureaucrats for helping them (their opponents) mobilize during elections.Footnote 45 Bureaucrats seek to avoid being assigned to work in a remote part of a district. One bureaucrat notes that “there are places you can be sent [if you get on the wrong side of the Bupati] from where you cannot return at all during the rainy season.”Footnote 46 Politicized promotions and transfers often force bureaucrats to navigate a complex landscape in which political influence overshadows their administrative competencies.
Actors in central agencies in Indonesia have taken steps to curtail individual politicians’ ability to interfere in bureaucrats’ careers. For example, the Civil Service Commission can overrule bureaucrats’ non-merit-based recruitment or promotions. The Ministry of Administrative and Bureaucratic Reform also forbade heads of districts and governors from immediately transferring bureaucrats upon taking office in 2016.Footnote 47 However, it does not adequately monitor such behavior.
7.1.3 What Motivates Politicians’ Use of Career Control Tools in Indonesia and India?
In Chapters 1 and 2, I argue that politicians have an incentive to use of these tools to facilitate either non-programmatic distribution or to capture public resources (these processes often go hand in hand). Evidence from India and Indonesia use career control tools for similar reasons.
Non-programmatic Distribution
As in Ghana, non-programmatic distribution is a significant feature of politics in India and Indonesia. It can occur either during or between election campaign periods. Between elections, politicians seek to direct public resources to certain individuals or communities. During campaigns, candidates often seek to distribute private benefits to individuals.
India has been described as a “patronage democracy” because of the high levels of political discretion involved in the implementation of public policy.Footnote 48 Indian politicians’ ability to direct public resources is partly embedded in formal institutions. In some states, politicians are reported to personally control 32 percent of rural development spending.Footnote 49 However, politicians seek to bolster their formal powers with additional tools. Their ability to transfer bureaucrats to different locations grants them informal discretion over the allocation of resources. Estimates suggest that this informal authority allows politicians to effectively gain control over an additional 49 percent of the development budget.Footnote 50 In other words, politicians gain control over a larger share of public resources through informal channels than they are allocated through formal institutions.
In Indonesia, as a legacy of the authoritarian period, bureaucrats exert significant discretionary authority over public resources.Footnote 51 Accordingly, politicians who want to distribute resources to their supporters have an incentive to either curry favor with bureaucrats or threaten them. Thus, politicians use transfers (or promises of transfers) to win over bureaucrats, thereby gaining greater discretion over public resources. Again, this highlights how politicians use transfers to enable them to engage in non-programmatic distribution.
During election campaigns, political candidates are under pressure from competitors to distribute private benefits to voters and club goods to communities. The Electoral Commission of India considers the “black money” that political candidates use to fund electoral handouts to be “the single biggest threat to the electoral process in the country.”Footnote 52 Electoral handouts are pervasive during election campaigns: survey evidence indicates that the vast majority of politicians at all levels feel pressure to distribute money and gifts to the public: 92% of surveyed national parliamentarians in India agreed that “peers feel pressure to provide gifts for citizens in a typical election;” 83% of state legislative assembly members, 86% of district council presidents, and 71% of block council presidents agreed with the statement.Footnote 53 Similarly, a qualitative study of party workers in Mumbai found that 19–64% of candidates’ campaign budgets were spent on gifts to voters or groups of voters.Footnote 54
Indonesia is similarly said to be “saturated in clientelism” during elections.Footnote 55 A comprehensive analysis of election campaigns across the country in the run-up to the 2014 general elections found that “patronage distribution is the central mode of political campaigning in Indonesian legislative election.”Footnote 56 The same study found that grassroots campaigns very rarely discussed programmatic politics – such as party brokers explaining a party’s welfare policies to constituents.Footnote 57
Political candidates engage in non-programmatic distribution such as handing out gifts to voters ahead of elections, distributing welfare services (such as medical check-ups and treatments), helping citizens access government-provided services, providing club goods to community groups, and introducing pork-barrel infrastructure projects.Footnote 58 During national elections, as many as one-third of the electorate in Indonesia receives a gift from at least one candidate.Footnote 59 High levels of poverty, the country’s open-list proportional representation (PR) system, and weak political parties encourage the distribution of campaign gifts.Footnote 60 Vote buying is less common in local than in national elections, although relatively large proportions of voters are still targeted with goods. One study finds that 13.6% of voters in regency elections and 11.3% in gubernatorial elections report receiving gifts.Footnote 61
Linking career control tools to the distribution of campaign gifts, past work establishes that promotions which coincide with the electoral cycle are more common in districts where clientelism is more entrenched, which suggests politicians use career control tools to extract bureaucratic loyalty to engage in non-programmatic distribution during elections.Footnote 62
Corruption and the Funding of Election Campaigns
The distribution of private benefits to voters – and, more generally, the need for face-to-face grassroots mobilization during elections – can impose significant costs on political candidates. As in Ghana, politicians in India and Indonesia often look to the state or to public sector processes, such as procurement, to capture funds.
India has the most expensive elections in the world: candidates spent an estimated $5 billion in the 2014 election, $8.7 billion during the 2019 election, and $16 billion in 2024.Footnote 63 There are formal legislative limits on how much candidates can spend on their campaigns,Footnote 64 and they are required to file election expenditure statements. However, in practice, candidates often underreport their expenditures and spend vastly more than these limits.Footnote 65
India is one of the few established democracies that does not provide public funding to parties. From 1969 to 1985, private firms were banned from donating to political parties. Together, these restrictions pushed the funding of political campaigns underground. While corporations were banned from donating, a quid pro quo was established between businesses and politicians: the latter provided regulatory favors or public contracts in return for donations.Footnote 66
Prior work demonstrates a close relationship between private firms and political candidates in contemporary elections.Footnote 67 This study finds that if there are close financial ties between private firms and politicians, this yields a 12 percent contraction in demand for cement during state elections. This is because in the months leading up to elections, building companies redirect their funds to party or candidate coffers; fewer resources are thus available for construction. While the authors do not directly link donations to public procurement outcomes, studies in other contexts suggest that in return for campaign funds, politicians are likely to award public contracts to firms that donate to their campaign.Footnote 68
High election costs can lead candidates to seek more nefarious ways to fund their campaigns. The extent to which corruption fuels politics is demonstrated by the results of a survey of Indian politicians: they reported that the most common source of funds for candidates were those “gained through corrupt activities/black money.”Footnote 69 Political candidates combine money from corrupt deals with private businesses or criminal networks with their personal funds and money from friends or other supporters.
Linking pressures to fund election campaigns and corruption with politicians’ use of career control tools, studies find that politicians’ threats of transfers can lead bureaucrats in India to engage in corruption on their behalf.Footnote 70 Bureaucrats fear that politicians will punish them for not engaging in corruption: “endemic political interference can lead to rent-seeking behavior even from honest officers, who might feel forced to comply with questionable demands from superiors for fear of being punished.”Footnote 71 Bureaucrats typically profit personally from engaging in rent seeking, but pass a significant part of what they gain onto politicians.Footnote 72
Indonesian political parties receive direct funding from the state. State subsidies covered 52 percent of the campaign activities of the Indonesian Democratic Party of Struggle (Partai Demokrasi Indonesia Perjuangan) in 1999.Footnote 73 However, following a presidential decree in 2005, state funding to national party offices is now “almost irrelevant.”Footnote 74 Total funds to political parties’ national headquarters were around $11 million in 2001–04 and dropped to just over $1 million in 2005–2009, and to less than $1 million in 2009–14.Footnote 75
Yet, electoral costs for parties and candidates have risen in Indonesia, partly due to the shift from closed- to open-list PR ahead of the general elections in 2004.Footnote 76 Direct elections for local government heads were also introduced in 2008, again increasing parties’ overall campaign costs.
The decline in state subsidies to central offices has also coincided with increases in the caps on donations that individuals and businesses can make to parties, increased auditing of party accounts, and stricter regulation on what parties can use public subsidies for. While political party finance appears to be well-regulated on paper, the monitoring of party accounts and sanctioning for financial malfeasance remains limited. The post-2005 reforms have arguably “shifted political fundraising even deeper into the shadowy realm of under-the-table donations, predatory use of state budgets, and oligarchic sponsorship.”Footnote 77 With increased campaign costs and reduced public funding to central party offices, grand corruption is increasingly linked to the funding of parties.
A major way in which political candidates cover their campaign costs is by awarding public contracts once they are in office. Politicians take bribes from contractors, or incumbent candidates can direct contracts to party financiers. Nearly three-quarters of experts (73%) surveyed in 2019 claimed that at least 40% of all big government contracts in their regions went to party financiers.Footnote 78 Most experts (69%) agreed that awarding government contracts to campaign supporters had become more prevalent over the last ten years.Footnote 79
Political parties also raise money for campaign activities by selling political candidacies. Once elected, candidates seek to recoup their nomination fees by using public contracting as discussed above, or by selling public sector positions to bureaucrats.Footnote 80 Trade in bureaucratic positions is worth a reported $1 billion a year.Footnote 81 More generally, parties expect candidates at all levels to largely self-fund their campaigns. Data from a nationwide survey of local legislative candidates shows that self-finance is the most important source of funding for about 90 percent of candidates in Indonesia.Footnote 82
Finally, parties and candidates fund campaign expenses and non-programmatic distribution by extracting state funds. Indeed, the state budget has been described as the “ultimate source” of funding for non-programmatic distribution ahead of national elections.Footnote 83 Local politicians distribute discretionary grant funding (dana hibah) to groups associated with the party or use social assistance and poverty alleviation programs to give money to voters. The expansion of dana hibah spending increased to more than a quarter of East Java’s total budget in 2014.Footnote 84
Does Electoral Competition Exacerbate or Remedy Bad Governance in Other Contexts?
Politicians should have greater incentives to use career control tools when local elections are very competitive. The channel through which high levels of electoral competition can hurt governance practices is through politicians’ greater reliance on non-programmatic distribution. There is evidence of a negative relationship between competition and the provision of public goods in India. The distribution of budgets and resources in India flows from State governments, through district governments, to blocs (panchayat samitis (PS)) and villages (gram panchayat (GP)). The GP is the most local level of government; they plan and distribute benefits directly to citizens and communities. The PS approves the budget and oversees the technicalities of GP projects. In line with the discussions above, such a system is said to provide the PS with “considerable discretionary power 
in project approvals and allocation of funds for different programs across GPs.”Footnote 85
To isolate the effect of competition on the allocation of public resources to GPs, prior work has leveraged the redistricting of constituencies in 2007.Footnote 86 Afterward, some GPs were in more competitive constituencies and others were in less competitive areas. Categorizing public spending into either public (i.e. local infrastructure) or private benefits a relevant study finds that increased electoral competition leads PSs to distribute more private benefits to politically aligned GPs in highly competitive constituencies. This has two important implications. First, this finding indicates that public resources are distributed according to political criteria: aligned local governments are allocated more resources. Second, it shows that electoral competition increases the distribution of non-programmatic benefits to garner electoral support. As in Ghana, it follows that politicians are likely to exert more pressure on bureaucrats to engage in non-programmatic distribution in electorally competitive constituencies.
Regarding the role of electoral competition in Indonesia, the evidence indicates that its transition from authoritarianism to democracy has increased politicians’ incentives to reward bureaucratic competence. At the national level, it is often challenging to evaluate the impact of introducing democratic elections on public sector hiring and post-hiring decisions, because there is typically a lack of high-quality data on hiring and promotion decisions under authoritarian rule. However, data on the career trajectories of bureaucrats from 1980 to 2015 in Indonesia makes such an analysis possible.
Bureaucrats’ career histories provide evidence that the transition to democracy increased the promotion premium for highly educated bureaucrats.Footnote 87 Education is more likely to predict promotion in the democratic period compared to the authoritarian period. This premium results in a 1–3 percentage point increase in the probability of promotion across different levels of education.Footnote 88 Bureaucrats with a postgraduate degree are twice as likely to be promoted under democracy as under authoritarian rule.Footnote 89 These results support the argument that democratic elections lead politicians to prioritize bureaucratic competence, as well as the conclusion that under authoritarian rule, bureaucrats’ promotions were partially based on other criteria. During the New Order regime, bureaucratic promotions are said to have been mainly guided by seniority and loyalty to superiors rather than performance.Footnote 90
The enhanced focus on bureaucratic competence has somewhat improved the delivery of public services. More generally, the shift to direct elections is correlated with the rise of health, education, and other welfare policies at the local level.Footnote 91 However, enhanced meritocracy also exists alongside intense politicization and high levels of corruption – most notably to deliver non-programmatic benefits to citizens, especially during election campaign periods.
The relationship between local electoral competition and governance in Indonesia is less studied. The deleterious effects of competition are suggested by one study that found more patronage hiring in districts that feature a strong opposition compared to those in which the ruling party is electorally secure.Footnote 92 Another study, which relied on experts’ perceptions of levels of clientelism, finds that while clientelistic practices are pervasive, they vary significantly by location.Footnote 93 The study attributes this variation to differences in the structure of the economy: it identifies less clientelism in districts with a more diverse economy. Economic diversification is argued to boost the “independence of citizens from ruling elites” and allow social forces (such as business elites) to constrain clientelistic practices.Footnote 94 However, these authors do not investigate whether there is a relationship between local electoral competition and clientelism across districts.
7.2 What Could Destabilize the “Career Control” Equilibrium?
The above discussions demonstrate that politicians beyond Ghana use career control tools to extract bureaucratic loyalty. Evidence from Ghana, India, and Indonesia suggests that politicians use such tools partly due to pressure to distribute funds outside of programmatic channels and fund their election campaigns. There is also evidence that higher levels of local electoral competition may encourage non-programmatic distribution, which increases politicians’ motivation to interfere in the work of bureaucrats. In Section 7.2, I discuss three sets of reforms that may improve governance in these contexts: (i) reforms to better separate the careers of politicians and bureaucrats, (ii) steps to encourage a greater transition to programmatic politics, and (iii) reforms to legislation that dictates the regulation of political finance.
7.2.1 Reforms to Separate the Careers of Politicians and Bureaucrats
My argument implies that greater career separation between politicians and bureaucrats can improve governance. Merit-based hiring is an important way to increase this separation. However, this book demonstrates that politicians can employ formal or informal tools to regain control over bureaucrats’ actions after they are hired. What reforms may help separate the careers of bureaucrats from politicians after they have been hired?
Bureaucratic transfers. To reduce the potential for politicization, governments could more closely control how long a bureaucrat spends in a particular post. Indeed, transferring IAS officers every sixteen months is unlikely to optimize bureaucratic efficiency in India.Footnote 95 Two reforms have been proposed there. The first is to introduce a minimum tenure for certain bureaucratic positions.Footnote 96 For example, bureaucrats could be prevented from being transferred until they have served at least two years in a position. The second proposal is to develop a “stability index” for specific posts to ensure that the average time in a position remains above a certain pre-determined average.Footnote 97 This proposal would limit politicians’ ability to rotate bureaucrats constantly to reduce the number of politicized transfers. However, Indian states have thus far resisted passing these reforms.
When considering reforms to limit political interference in the posting of bureaucrats, a balance must be struck between rigidity and flexibility. For example, imposing fixed lengths of time for posts is likely to generate significant costs, which could demotivate bureaucrats assigned to particular posts and make it difficult for governing elites to respond to genuine vacancies and changes in circumstances. Furthermore, extended fixed-term postings may encourage bureaucrats to become embedded in local networks, which may encourage corruption.Footnote 98
Promotions. As discussed earlier, local bureaucrats’ promotions are politicized in Indonesia but not Ghana. In Ghana, panels of bureaucrats (rather than politicians) make promotion decisions, while one or two individuals decide transfers. Promotion timelines are also relatively rigid in Ghana: bureaucrats expect to be promoted after a fixed number of years, which provides less room for politicization. However, a significant downside is that rigidity does not tightly link promotions to performance, which can discourage effort. That said, an overtly politicized promotion process can also discourage effort. Accordingly, as with transfers, a balance must be struck between rigidity and flexibility in promotions. A crucial element of flexibility is ensuring that the promotion system allows exceptionally hard-working and high-performing bureaucrats to be promoted early.
Work assignments. As discussed above, politicians can manipulate bureaucrats’ careers by interfering in the type of work they are assigned to do. Reforms to limit this control are likely to be context specific, but generally entail ensuring that bureaucrats’ peers (rather than politicians) monitor their day-to-day work portfolios and performance.
Tenure. This book assumes that bureaucrats have tenure of office and thus politicians cannot threaten to fire them; however, this is not always the case. Indeed, merit-based hiring and tenure do not necessarily coincide, especially in developing countries.Footnote 99 Securing tenure for bureaucrats is an additional essential step to separating the careers of politicians and bureaucrats.
Administration. Finally, Establishing or empowering centralized agencies that administer the public sector can better insulate bureaucrats’ careers from political interference. For example, Japan’s National Personnel Authority (NPA) is an autonomous and non-political administrative agency. The National Public Employees Law of 1947 stipulates its remit. Among other responsibilities, the NPA is in charge of administering the general public service examinations, recommending bureaucratic salary levels to parliament, administering benefits, managing training, and reviewing disciplinary measures.Footnote 100 This agency has been key in securing bureaucratic efficiency and good governance practices in Japan.
Beyond the formal structure of the public administration, reducing politicians’ incentives to engage in non-programmatic distribution can shift the governance equilibrium, the topic to which I now turn.
7.2.2 Transitioning to Programmatic Politics
Politicians’ quest to engage in non-programmatic distribution incentivizes them to use career control tools to solicit bureaucratic loyalty. When politicians prioritize programmatic distribution via universalistic policies, they will be more likely to favor bureaucratic competence and limit their interference in administrative processes. Both supply- and demand-side factors can motivate politicians to prioritize programmatic distribution. The most important demand-side change is an increase in voters’ wealth. Important supply-side changes include alterations in social structures, such as those resulting from population growth and urbanization, as well as political changes, for instance, when national governments directly provide universal welfare benefits to citizens.
The most consequential demand-side change that is likely to reduce the salience of non-programmatic politics is an increase in constituents’ average wealth.Footnote 101 As voters become more economically secure, particularistic benefits become less valuable to them. Increased opportunities for private sector employment can foster private sector growth and economic security. While individual politicians can attempt to increase the value of the discretionary benefits they offer voters, at some point it is likely to be strategically beneficial for them to provide universal benefits.
Structural changes in the UK and U.S. electorates caused by industrialization and economic growth are said to have “eventually spelled the demise of clientelism.”Footnote 102 Clientelism is thought to have persisted for longer in the United States because industrial growth still left many voters economically vulnerable. The extension of public welfare following the New Deal helped reduce poverty and supported greater income equality,Footnote 103 and the increased welfare provisions reduced the attractiveness of private inducements from politicians via party machines.
Survey evidence suggests wealthier voters are less inclined to vote for politicians who engage in non-programmatic distribution.Footnote 104 These preferences also influence politicians’ behavior: politicians representing districts with a large presence of middle-class voters and high levels of electoral competition are less likely to engage in clientelism.Footnote 105
Individuals’ average wealth, as well as the distribution of middle-class voters within or across districts, determine the extent to which demand-side changes affect politicians’ strategic calculations. Middle-class voters need to be geographically concentrated within relevant political borders – districts, municipalities, constituencies, etc. – to alter politicians’ incentives.Footnote 106 When middle-class or wealthy voters are geographically dispersed, they do not make up the majority of voters and politicians may not feel electoral pressure to transition to programmatic distribution.Footnote 107 Frustrated by the lack of political parties or politicians offering programmatic alternatives, wealthy voters may decide not to vote.Footnote 108
Supply-side factors that alter the structure of society can reduce the ease with which politicians can reach voters and provide goods to them. One important social change is urbanization. Economic growth in the UK increased urbanization; both reduced politicians’ ability to target individuals with discretionary benefits. Voters became more anonymous within cities: as the size of urban constituencies increased, politicians began using different tactics when campaigning for votes. Indeed, it has been argued that “the cost of electioneering on the old lines would be quite problematic”Footnote 109 – a reference to a reliance on party brokers and the distribution of bribes and intimidation.
Another important supply-side change that decreases individual politicians’ incentives to engage in clientelism is the introduction of national, universalistic welfare programs. Contemporary examples of the introduction of such programs include Bolsa Familia in Brazil and Progresa/Oportunidades in Mexico. These programs led to a reduction in clientelism.Footnote 110 Yet the introduction of national welfare programs does not always reduce clientelism, since local politicians may capture them. The governments’ decisions to centrally administer both programs put national officials in charge of selecting beneficiaries and distributing program benefits.Footnote 111 Since both programs had widespread coverage of target populations, politicians lost the ability to reward or punish voters with valuable goods.
The introduction of universalistic welfare programs influences both supply and demand factors. Regarding supply, such programs are likely to divert some financial resources from local politicians. On the demand side, as individuals receive benefits, their economic security should increase, reducing their expectations of obtaining benefits from local politicians. If the introduction of centralized welfare programs holds an important answer to the question of how to improve governance, a key secondary issue is the conditions under which national leaders and their cabinets have incentives to introduce and insulate such programs.
7.2.3 Reforming Political Finance Legislation
My findings suggest that stronger regulation of political finance, along with introducing or expanding public subsidies to parties, may reduce political interference in public administration. Politicians are partly motivated to use career control tools to solicit bureaucratic loyalty to extract public money or collect kickbacks from private companies to fund their election campaigns. The middle-income democracies I discuss in Chapter 7 either do not fund political parties (India and Ghana) or provide negligible funds to national party organizations (Indonesia). These countries have no practical limits on how much candidates or parties can spend during elections. While India has legal spending caps, they are not enforced and are routinely violated. Due to high election costs and limited (or no) state funding, parties and candidates seek alternative funding sources.
Many countries provide public subsidies to political parties.Footnote 112 Public subsidies are often parties’ main source of funding: in some European countries, they constitute 85 percent of parties’ total income.Footnote 113 Theoretically, introducing (or increasing) public funding to parties can reduce corruption by decreasing the marginal value of private donations.Footnote 114 A reduced dependence on funds from private actors should then reduce politicians’ incentives to engage in quid pro quo deals with private actors.Footnote 115 There is mixed empirical evidence to support the claim that public subsidies decrease corruption.Footnote 116 However, recent work using global longitudinal data finds overall support for the hypothesis that public funding reduces corruption.Footnote 117
Political finance reform can also entail more strongly enforcing campaign spending limits; nearly half (46%) of all countries limit what candidates can spend during election campaigns.Footnote 118 However, there are regional disparities: 67% of European countries (but only 34% of African countries) impose spending caps on candidates.Footnote 119 Stricter limits on how much parties are allowed to spend can reduce corruption by decreasing the cost of elections, thereby weakening candidates’ incentives to extract public resources and engage in corruption.Footnote 120 Stricter spending regulations also encourage the diversification of political candidates.Footnote 121
The UK’s adoption of the Corrupt and Illegal Practices Act (1883) significantly reduced election-related corruption.Footnote 122 Sir Henry James, the Liberal attorney general who introduced the bill, called excessive electoral spending the “very father of corruption.”Footnote 123 The law imposed harsh punishments on candidates who did not comply with the law, including significant fines, politicians being unseated and banned from running for office in the future, and imprisonment.Footnote 124 It had enormous and immediate impacts: between the 1880 and 1885 elections, total campaign expenditure fell by 40 percent.Footnote 125 This coincided with a move away from non-programmatic distribution; candidates were forced to rely more heavily on campaign messages via speeches, which were printed in newspapers.
The case of India demonstrates that to effectively curb corruption, spending limits must be strictly enforced. In Brazil, candidates who file irregular accounts can lose their seat and be banned from running for office in the future; parties can lose access to public subsidies and be prosecuted court.Footnote 126
If greater political finance regulations can improve the functioning of public bureaucracies, a key question is where the pressures for such reforms come from. In general, it is challenging to theorize reform adoption, because there are numerous pathways to reform, and the specifics of the reform options vary. For example, what motivates politicians to introduce party subsidies may differ from what prompts them to limit campaign spending. One potential model is to conceive of reforms occurring when political elites’ opportunities and motivations coincide.
Democratic transitions provide an important window of opportunity to adopt reforms. For example, many post-Soviet countries introduced public subsidies when they transitioned to democracy, arguably motivated by parties’ need for revenue.Footnote 127 The number of institutional veto players also shapes the opportunities for reform, as do international factors such as the actions of neighboring countries. There is evidence of policy diffusion in political finance regulation across countries.Footnote 128
Parties are likely motivated to adopt reforms when it gives them an electoral advantage or helps mitigate electoral loses.Footnote 129 Accordingly, corruption scandals can be an important impetus for reforms, especially when the public discourse connects corruption to party finance.Footnote 130 For example, revelations about the UK Conservative Party’s reliance on foreign donors led the Labour Party to include political finance reforms in its 1997 campaign manifesto.Footnote 131 This eventually led to an overhaul of the country’s political finance regime with the adoption of the Political Parties, Elections and Referendums Act in 2000.
7.3 Avenues for Future Work
The results I present in this book, along with the discussions in Chapter 7, suggest at least three important avenues for further research. First, future work could consider the incentives and ability of politicians and bureaucrats to efficiently deliver public services and improve governance. A common argument is that low state capacity constrains the electoral strategies that politicians adopt in developing countries.Footnote 132 There is, of course, some truth to this argument: resource constraints do impact states’ ability to deliver goods to citizens.Footnote 133 However, this book establishes that politicians’ actions often cause weak state capacity. In diverse contexts, politicians have been shown to have strategic incentives to reduce state capacity.Footnote 134 They also create, or allow the creation of, an excessively complex bureaucratic or legislative structure.Footnote 135 This is because when administrative procedures are burdensome and error prone, this presents an opportunity for re-election-seeking politicians to engage in case work on behalf of their constituents. Politicians occupy a unique position as the “monopoly suppliers of bureaucratic ‘unsticking’ services.”Footnote 136 Therefore, a fruitful avenue for future research entails investigating the conditions under which politicians have incentives to strengthen or weaken, streamline or complicate the running of state agencies, and forgoing the assumption that politicians rely on non-programmatic distribution due to low state capacity.
Second, and related, conceptualizing governance as the result of the actions and expectations of politicians, bureaucrats, and voters reveals that more research is needed to understand when (and why) politicians have an incentive to abandon clientelistic politics. This is likely linked to voters’ demands. I highlight that local electoral competition creates incentives for politicians to interfere in administrative processes and engage in corruption. However, prior work demonstrates that local electoral competition, combined with a large middle-class presence, encourages politicians to move away from clientelistic practices.Footnote 137 As noted above, other research suggests that the complexity and diversity of the local economy play a crucial role in determining the salience of non-programmatic distribution.Footnote 138 Given the fundamental ways in which politicians’ reliance on non-programmatic distribution influences governance, this is an important avenue for future research.
A third avenue for future work entails deepening our understanding of what policies can successfully separate politicians’ and bureaucrats’ careers.Footnote 139 As discussed earlier, such research must go beyond a focus on meritocratic hiring and tenure protection to consider how to protect bureaucrats’ careers over the long term. Successful reforms may seem small and technical, such as the composition and chairs of promotion committees, but may be highly consequential. Solutions to protect bureaucrats from corruption include making it more difficult for them to interfere in public procurement processes, such as introducing and strengthening electronic procurement systems.Footnote 140 In-depth qualitative interviews and participant observation in administrative offices may help identify other reforms that grant bureaucrats more control and autonomy over administrative processes. An important takeaway from this book is that bureaucrats often deeply desire to serve the public interest.Footnote 141 Engaging with bureaucrats to better understand how to help them fulfil this mission has the potential to improve governance outcomes worldwide.
Open access