A Theory of Political Organization

Joseph Luna∗

April 19, 2015

Abstract

Party finance shapes political organization. Drawing from fieldwork across Ghana, I identify four sets of actors—politicians, bureaucrats, construction contractors and political–party chairs—who collude to extract financial resources from the state. These resources are extracted through the public–procurement process and legitimized by the bureaucrat’s monopoly over expertise. Party chairs demand rents from political incumbents and aspirants in exchange for mobilizing party operatives during the electoral season. Politicians and bureaucrats oversee procurement committees across the country, and award contracts to favored contractors who return kickbacks. Failure to pay rents unleashes a chair’s credible threats against a politician. Ultimately, this system diminishes the quality of development. This paper details a political equilibrium and the practices that sustain it.

∗Author email: [email protected]. I express tremendous gratitude to Robert Bates, Daniel Car- penter, Arthur Spirling, Lucie White, Kojo Asante, Ebenezer Oﬀei–Ansah, Mohammed Awal, Jonathan Phillips, Leah Rosenzweig, Mauricio Fernandez Duque, Matthew Stephenson, Daniel Franklin, Anna Pers- son, Rachel Sigman, Staﬀan Lindberg and colleagues at Northwestern, Harvard and MPSA for fruitful discussions on previous versions of this work. I am grateful to the hundreds of Ghanaian public servants and private individuals who contributed their time and warm hospitality, as well as the support of the National Science Foundation and Harvard University’s Frederick Sheldon Fellowship. This research is authorized under Harvard IRB Protocols F21566: 101–104. This piece is adapted from my dissertation–in–progress. Please do not circulate.

1 I. The Importance of Political Organization

“Organization is the road to political power, but it is also the foundation of political stability and thus the precondition of political liberty. The vacuum of power and authority which exists in so many modernizing countries may be ﬁlled temporarily by charismatic leadership or by military force. But it can be ﬁlled permanently only by political organization. Either the established elites compete among themselves to organize the masses or dissident elites organize them to overthrow the system. In the modernizing world, he controls the future who organizes its politics.” —Samuel Huntington, Political Order in Changing Societies

Without political organization, the demands of citizens for public goods and services would be difficult to realize. Political parties are the central units of a state’s overall political organization, but parties do not exist within a vacuum: bureaucracies, private businesses, civil society and traditional leaderships interact with parties in the governance of a state. Such entities can provide popular support for political parties while also chan- neling the demands of citizens. But most parties are not merely receivers of support and citizen demands. Most parties desire to win political offices. To do so, they must raise money, and how that money is raised defines a party’s relationship with the other political entities and, ultimately, how it rules. Early political sociology explored the nature of political parties. For Michels [1915], political parties were paradoxical: despite their democratic beginnings, they developed specialized bureaucracies that doomed them to become oligarchical organizations (see also Ostrogorski [1895]). Weber [1919] noted the traditional and charismatic characteristics of political parties and the role parties play in inserting partisan ideology into state systems.1 While Michels’ [1915] “iron law of oligarchy” is found in numerous examples, Lipset et al. [1956] detail conditions underlining the exceptional case of the International Typographical Union. Political parties can deliver the goods and services that the state is failing to provide. Riordon [1905] describes New York’s Tammany Hall machine. Such a machine represented immigrant interests and ensured that supporters were rewarded with patronage jobs in the civil service or with construction contracts. The machine’s leaders controlled important

1See Mudge & Chen [2014] for an overview of the early political sociology of parties. These authors also describe a resurgence in the political sociology of parties following a hiatus that started in the late 1960s.

2 political offices that allowed for the funneling of contracts—or, “honest graft”. I argue below that similar dynamics—goods provision and political enforcement—exist in my case country. Rational–choice theorists examine parties from the perspective of office–seeking politicians. Aldrich [1995, 2011] argues that office–seeking politicians form parties to solve collective–action problems, namely electoral mobilization and the formation of long leg- islative coalitions. Being attached to a party cues voters to a particular “brand name” that can be useful in deciding one’s vote. A long coalition allows groups of politicians to pass legislation that serves their interests (see also Weingast [1979]). Moving away from rational–choice, candidate–centered theories of parties, Bawn et al [2012] theorize that parties in the United States are better understood as coalitions of interest groups seeking to capture the government for their own purposes. These authors assert that interest groups influence parties most at the nomination stage, a stage which most voters ignore (see also Cohen et al [2008]). As evidence, the authors point to shifts in party position as a result of interest–group pressures; notable policy examples include the pivots towards civil–rights protections (Democrats) and anti-abortion legislation (Republicans). Beyond candidate–centered and interest–group–centered theories of political parties lies another issue. Whether parties and politicians are most concerned with providing tangi- ble goods to their citizens or whether they serve business interests, methods of financing campaigns shape ultimate actions. In the American context, Heberlig & Larson [2012] describe the upending of the seniority advancement system (for committee chair positions) in the US House of Representatives in favor of a pecuniary system where candidates who contribute the most to party campaign committees from their own political–action committees (PACs) receive their top committee choices.2 Since the 1990s, candidate fundraising through PACs for this purpose has increased substantially as incumbents work harder to rally their donors. Whom a politician owes, defines how she will behave. The theory I derive centers on political finance. While its framework borrows from the sociology described above, the interactions between individual actors within that framework can be thought of from a rational–choice perspective. This theory seeks to answer the question of “who organizes politics?”, which also entails identifying who enforces the rules to maintain the equilibrium. I will identify how politicians and parties are financed, and how that process is legitimized

2These donations can help the party receive a majority in the House, which is a necessary condition for awarding chair positions.

3 to prevent prosecution and voter unrest. Policy implications will also be discussed. Who controls the ﬁnancing of politics, organizes it.

II. The Case of Ghana

The theory below is derived from ﬁeldwork conducted over several years in Ghana. A developing country in West Africa, Ghana provides insights into political ﬁnance and organization that are relevant for other developing countries.3

A. Political Background

Formerly a British colony, Ghana established a democratic constitution in 1992 and maintains a strong–presidential system of government.4 Administratively, the country comprises 10 regions, which are subdivided into 216 districts. There are numerous entry points into politics for Ghanaians. According to Schlesinger [1966], political ambitions are shaped by the structure of available offices. In Ghana, the lowest political offices are unit committees, which correspond to ward aldermen (FES 2010). Just above the unit committees are District Assembly members, who form the deliberative body at the district level.5 The highest office in a district is the District Chief Executive (DCE), similar to a mayor. The DCE is appointed by the President and serves at the President’s pleasure—her term is not fixed. This person is almost always a member of the President’s party. As a result, the governing party holds the highest office in each district, even in areas that are opposition strongholds. The DCE’s official task is to ensure that the government’s policies are implemented in the district, and she also oversees the allocation of local contracts and the monitoring of projects. Beyond the district, the local Member of Parliament is the other significant political official. Members are elected from 275 single–member constituencies, and they serve fixed four–year terms.6 According to the 1992 Constitution, at least half of the President’s Cab- inet must be comprised of MPs. Presidents are elected from a single national constituency,

3I would further argue that there are numerous parallels between political–ﬁnance behavior seen in Ghana and in the United States. 4Between being a colony and a democratic state, Ghana was ruled by various military and civilian governments. 5In each district assembly, approximately 70% of members are elected by the district’s citizens, while the rest are government appointees. 6Most rural constituencies overlap directly with the district.

4 and also serve fixed four–year terms, coterminous with MPs. Since 1992, Ghana has had four Presidents: all possessed political experience prior to becoming President. Similarly, most vice presidents have had political or civil-service experience, with the exception of Alhaji Aliu Mahama, vice president from 2000–2008 with the New Patriotic Party (NPP), whose members had been in opposition for decades prior to 2000. Aliu Mahama never served in any political position, but was a prominent contractor of the Northern Region and a key NPP financier. In recent years, academics have served as Vice President (John Atta Mills, Kwesi Amissah-Arthur) and as Vice Presidential candidates (Mahumudu Bawu- mia). 7 Two parties dominate Ghanaian politics today. The social–democratic National Demo- cratic Congress (NDC) is the current ruling party, capturing both the presidency and parliament. The NDC derives much of its electoral strength from the Volta Region, traditional home to the Ewe ethnic group, as well as the Muslim–majority northern regions of Ghana. The business–oriented New Patriotic Party (NPP) is the current opposition, and last held the presidency and parliament from 2000–2008. The NPP’s electoral strongholds are the Ashanti and Eastern Regions, which are the traditional homelands of the Akan ethnic groups.8 The President of Ghana wields numerous appointment powers. In addition to appoint- ing ministers and deputy ministers, the president appoints all 10 Regional Ministers and all 216 District Chief Executives. However, the President makes these choices in consultation with the ruling party’s executives, and it is not clear that the President dominates that process. These appointed officials can be replaced at any time, so they strive to satisfy the President and the ruling party. From a bureaucratic standpoint, Ghana recently embraced decentralization. While there is a centralized policy–formulating bureaucracy in Accra, each region and district has its own policy–implementing bureaucracy.9 The District Coordinating Director (DCD)

7It is possible that academics have recently found favor in politics because they garner respect from Ghanaian voters but lack political bases of their own, which may give party executives more leverage over them. 8It is technically possible for the presidency to be held by one party and parliament, by another. Recently, more constituencies have voted ‘skirt–and–blouse’, choosing one party’s candidate for president and the other party’s, for parliamentarian. 9The centralized agencies include such groups as the Ministry of Health, Ministry of Education, etc. The decentralized bureaucratic apparatus is overseen by the Ministry of Local Government & Rural Development as well as the Local Government Service Secretariat. Some agencies, such as the Ministry of Food & Agriculture have had a decentralized presence, but those oﬃcers are currently being folded into the local– government structure of each region and district.

5 is the administrative head of the district. Typically a long–serving veteran of the public service, she oversees the activities of every bureaucratic sector in the district. The District Finance Officer manages the financial and tax–raising mechanisms of the district. The District Planning Officer assesses the district’s overall development needs. The District Works Engineer oversees public construction and procurement. The District Budget Officer drafts budget planning and allocation reports. Program areas such as education, health and agriculture have their own directors, but they tend not to work in the district’s central office.10 Bureaucrats are appointed via the Local Government Service Secretariat in Accra. Their training is extensive but tends to vary, with a wide range of fields of study and levels of graduate education.

B. Sample & Methods

Ten months of principal ﬁeldwork were conducted in Ghana between September 2013 and July 2014.11 Ten districts were chosen for the main study, and one pilot district was studied early on to prepare for the others. Each region had one representative district in the main study. Sampling took several criteria into consideration. Every district has a DCE selected by the ruling party, but MPs can come from the NDC or NPP. I ensured balance between districts where both major politicians were from the NDC and districts where there was a divide. Secondly, DCEs can serve multiple terms, so districts were selected in which the DCE had been re-appointed and those where the DCE was new. It was also important to account for DCE ambitions, so sampling considered whether the sitting DCE (new or incumbent) had run in either the primary or general election for Parliament in the prior election (2012). Finally, I ensured that at least one of the 10 districts was newly created, and that three were more urban, which is proportional to the rural–urban composition of the 216 districts in total.12 Ethnographic methods formed the core of this research. I lived in each district’s capital, and most days were spent at the district–assembly buildings, where I became familiar to the DCE, bureaucratic oﬃcers and other local notables. For the most senior politi-

10For more on district assemblies, please see Friedrich Ebert Stiftung’s 2010 “A Guide to District Assem- blies in Ghana.” 11This project builds oﬀ of ﬁeldwork conducted over several trips starting in 2008; in total, more than two years were spent conducting research in Ghana and cultivating relationships with politicians and various public agencies. 12Forty–six new districts were inaugurated in 2012. With each new census, the government of the day has the right to carve out new districts.

6 cians and officers, I conducted in–depth, semi–structured interviews, taking between one to three hours per interview. My residence in each district capital allowed for numerous informal discussions, whether at their residence, a drinking spot or in their vehicles. I also participated in numerous assembly meetings and project inspections. Other officers—namely, planners, engineers, some junior officers, etc.—were directly administered a survey by me.13 This survey included list and conjoint experiments that tested the criteria each officer would use in selecting a contractor (Glynn 2013; Hainmueller et al 2013). The results of these experiments are presented in separate work. Beyond politicians and bureaucrats, numerous interviews and focus groups were conducted with former DCEs, party officials, contractors and chiefs. Contract data was gathered in each district. While each district is nominally supposed to report contracts to the Ministry for Local Government & Rural Development, in practice this does not happen. Such data is rarely publicly available.14 The quality of the data differs significantly between districts. While in Accra, I obtained data on the universe of qualified roads and works contractors, and could thus verify whether contractors operating in a district had valid qualifications. The allocation process for government contracts is prescribed under the Public Pro- curement Act of 2003. This act details the monetary thresholds that a district cannot exceed in awarding contracts.15 Further, this act also describes the process for advertising contract availability in national media outlets and ensuring fair bidding. The act notes that, except in emergency cases, each project must have at least three bidders to ensure competitiveness. In practice, many provisions of this act are ignored. Methodologically, my work studies corruption up close (see also Blundo & Olivier de Sardan 2006). This is not a study of ‘perceptions’ of corruption, but of actual corrupt processes. This is not an experimental study of petty corruption, such as bribing the police officer, or a study of grand corruption at the presidential level. Rather, I consider this ‘mid– level’ corruption to be especially dangerous because it is not large and concentrated enough to be noticed by the media, but the amounts and projects involved significantly impact the quality of public services provided to citizens.16 13In total, 83 of these officers were interviewed across the 10 main districts. 14However, most districts did provide their 2012 Annual Action Plans, which for many includes basic data on contract winners and bid amounts, but little detail on the actual award process. 15Districts are not to exceed 50,000 Ghana cedis per project contract. Most districts simply bypass that threshold, but there are also numerous contracts valued around 49,000 Ghana cedis. Further work will explore this data. 16A final note on methods: I am frequently asked, ‘How did you get people to talk about corruption?’

7 III. From Practice into Theory

“Since the politician is also always a citizen, it is never possible to abstract him as pure oﬃceseeker, in the same degree to which we can isolate the business- man or the doctor. The politician’s other interests are always involved in his decisions as politician.” —Joseph Schlesinger, Ambition and Politics

The rational, pecuniary interactions between four critical players underpin political organization in Ghana. These four players are the politicians, bureaucrats, contractors and party chairs. Their individual preferences are actualized into concrete actions through the distribution of project contracts. These contracts are primarily funded from two sources: central government revenues and donor revenues. Central–government revenues are dis- bursed through the District Assemblies Common Fund (DACF).17 Donor revenues are centralized and distributed to districts under the District Development Facility (DDF).18

A. Preferences

Preferences guide actions. For politicians, scholars posit numerous theories for what mo- tivates their actions, ranging from the reelection incentive, desire for ﬁnancial gain and aspirations to help one’s community (see Mayhew 1974; Fenno 1978; Fenno 1990; Lessig 2011). In Ghana, many politicians asserted that they ran to serve their people.

“I didn’t regret serving the people. I served from my heart. Those of us in the ﬁrst Kufuor term—we served from our hearts. I’m not only serving my own people.” Former DCE, New Patriotic Party, Forest Zone

However, political oﬃce is prestigious. It commands the respect of one’s fellow citizens— but it also exposes the politician to numerous requests for assistance (see Lindberg 2010).

This project was enhanced by living in each district and spending signiﬁcant amounts of time interacting with people. Familiarity turned to trust; interviews became confessions. For a typical public servant and politician in my sample, almost all were aware of the corruption involved in the procurement process and felt varying degrees of guilt or helplessness. For them, a foreign researcher committed to their district provided an opportunity for catharsis—and many took it. While each interview is treated academically by this researcher, many respondents were relieved to have someone listen to their life stories without judgment. 17District allocations are determined by a formula, with allowance for ﬂuctuations. Part of the DACF is reserved for the MP(s) representing that area, and these MP funds will be addressed below. 18Districts can also raise internal funds through local taxation, but in practice these amounts are minimal in all but the most urban districts.

8 “That one is terrible! They come in as early as 4AM on phone. Some come to parliament with social issues, school fees, funerals, marriage, outdoorings, harvests. All manners of people: some want employment with police and security. They look at you as their lord. If you don’t pick it, there is a problem.” Member of Parliament, New Patriotic Party, Forest Zone

For a Ghanaian politician, reelection entails satisfying these material needs of the people in addition to credit claiming. Politicians who fail to provide such financial assistance and who do not maintain good relations with their local party executives face unrest or a strong primary challenger. In this theory, I assume that politicians prefer to win office or stay in office. (cf. Schlesinger 1966). Many DCEs desire to be elected to Parliament. According to my calculations, in the 2012 parliamentary election cycle, approximately 52 DCEs contested for parliament, at either the primary– or general–election stage. Of these, 13 were elected to Parliament. Ghanaians provided similar reasons for joining the public service. Many wanted to “serve their country”; others were simply looking for a job after university and wanted to apply their skills. The bureaucracy has long been a respected position in Ghana, and, like politicians, bureaucrats are expected to respond to the requests of friends and extended family (see also Achebe [1960]). They, too, have expectations to meet.

“As for public servants, they are not all that different from politicians. They want financial gain, a nice house, a nice car, etc. People expect that. There are cultural demands. You, for instance, are getting your PhD at Harvard. People will have expectations about the type of life you will be leading.” Planning Officer, District C

A large portion of bureaucrats interviewed mentioned retirement as their top concern. Bureaucrats report that they are expected to build a house to which they can live, once they reach the retirement age of 60. It is considered shameful if a bureaucrat has not planned for his or her accommodations.

“First and foremost, in Ghana, we think of housing and start preparing for that. Civil servants—using money to build a house and we get loans from bank. Stress will be on you if you build after you retire. Some have insurance, etc. Foremost is to get a house. Second is to educate your children. For me,

9 when I started work, my children were small and you just buy small to build your house. Stay focused on plan.” Food & Agriculture Oﬃcer, District D

Construction is a large industry in developing countries. The descriptor developing im- plies that these countries need such infrastructure as roads, schools and clinics. In Ghana, it is relatively straightforward to become a contractor and start one’s own ﬁrm. While, on paper, there are technical requirements that must be met, in practice, these requirements are rarely enforced. Given the widespread need for public infrastructure, being a contractor can be lucrative—provided one has the right connections. Most contractors, however, are more concerned with simply turning a proﬁt.

“You see, Joe, in this country anyone can be a contractor. You don‘t have to go to university, you don‘t have to be Bill Gates. All that you need to know is that a politician needs money. That contractor may not be very intelligent. But he knows real politics.”Contractor, District C

In Ghana, political–party chairs are powerful. They control access to the “brand name” of the political party, which candidates need to win votes. From a logistical standpoint, party chairs control the “foot soldiers”, party loyalists who are mobilized to rally votes during an election. Party chairs want their preferred candidates to win elections or be appointed as DCEs, and ultimately they are interested in their party capturing the presidency. Party chairs have numerous responsibilities, as well:

“Because my party is in power, I help my people. When I win the election, I‘ll help people with funerals, weddings, etc [...] As for my relationship with contractors, when jobs come up here, I help them so they have something to eat.” Regional Party Chair, National Democratic Congress

These four players have individual preferences, but these preferences have something in common. Achieving these goals requires money. Individually, few members of each player set has the individual wealth to achieve their goals. Ghana’s private sector has until recently grown robustly, but it is not as signiﬁcant an economic player as the central government. In a developing country like Ghana, it is the central government that ﬁnances development. This presents opportunity. For every development policy that is actualized into a school block or road or clinic comes a process—procurement—by which a contract is

10 awarded to a builder. A large sum of money is being transferred from the state to a local agent. These players together rig the procurement process to extract rents—and they do so in a way that maintains legality.

B. Actions

“Politicians need money to do campaign. I am a contractor. He advertises projects. Politician gets project and his money for campaign.” Contractor, District C

Political ambitions initiate the equilibrium. Both incumbents and aspirants require money to campaign for elected or appointed oﬃce. For Parliamentary primaries, incumbents and aspirants must often purchase the votes of party executives to reach the general election.

“Most don’t fund campaign alone: friends, relatives, contractors—not just loans. There are expectations. If you want to go to Parliament, if from a district, most rich people will assist you. If you can become Minister, you can help them get contracts.” Regional Party Secretary, National Democratic Congress

Campaigns, especially with vote buying, can be expensive. In January 2014, the Majority Leader reported that each MP in the House owes at least 30,000 Ghana cedis (then about US$10,000) to private banks for their campaigns.19 Aspirants can borrow money to buy votes to win elections, but they must pay that money back. Member of Parliament salaries are around 70,000 Ghana cedis annually.20. DCEs receive less, but politicians must still satisfy constituents’ demands. Contract kickbacks provide a means for earning extra money.

“Ten years ago, no one took 10 percent. This changed when politicians realized they could only make money in construction—you see him turn his brother into contractor. But we technicians are locked out. We try to get closer to MPs; the MPs struggled to campaign, and you must give to them.” Contractor, District H 19http://www.myjoyonline.com/politics/2014/january-19th/each-mp-owes-ghc300000-majority-leader.php 20Author interview with MP, District E, April 2014

11 Access to public contracts is vital for Ghana’s contractors. In theory, the public– procurement process should ensure that qualiﬁed contractors bidding at appropriate prices are chosen. However, this process introduces discretion. The contractor with the lowest bid price to build a school, for example, may not get chosen, because having the lowest price might imply lower quality. Politicians and bureaucrats must assess the quality of a contractor to ensure they receive value for money. This places politicians and bureaucrats in a position of power, and contractors must win their favor before winning a contract.

“Before you tender, make sure you have all arrangements set. Here, I get in touch with MCE, DPO, DBO, etc. They have to know you well. If there is a tender, and you don’t do your homework, you lose.” Contractor, District H

“Contractors don’t always involve themselves with the executive, but they have parties they support. Namely ﬁnancial support, such as during campaigns. You wash my back, I wash yours. If they want contracts, it is easy for them to get through.” Constituency Party Secretary, New Patriotic Party

A contractor will promise politicians and bureaucrats a kickback—which is typically phrased as a “thank you”—from the awarded amount. Failure to provide this kickback will blacklist contractors, preventing them from winning future contracts. Bureaucrats play a critical role in this process. Each district has several specialized bureaucrats: an engineer, a finance officer, a planner, in addition to the coordinating director who oversees them. Bureaucrats possess an advantage over politicians. They are the expert technicians in policy implementation. A politician may publicly support a project, but a bureaucrat’s support adds legitimacy. In the procurement process, the bureaucrat’s imprimatur validates the process that the best, most cost–effective contractor was chosen, even if the reality is different. This shields politicians, bureaucrats and contractors from a zealous auditor, whether from the central government or an external donor. Consonant with principal–agent theory, the cost for these outsiders to monitor local bureaucrats is too high.

“Politicians have friction because they know they have a maximum term, but a public servant has until the age of 60. Politician thinks that when he comes he wants to make the best of his years. This leads to corrupt practices. Some lack an understanding of the rules and regulations—they want public servant

12 to set aside rules, regulations and laws to do what is expedient. We have a procurement law (Act 663). Politicians would want you to initiate a development project without going through process. You must use diplomacy: let him un- derstand that you are going through process in his interest. You will often get sympathizers for the politician who will frustrate you. Bureaucrats don’t work in their hometowns, and the politician may think you are not for development in his area.” District Coordinating Director, District I

It is important for a bureaucrat to always remember to go through the procurement process in the politician’s interest. How does a bureaucrat actually facilitate this process? Figure 1 presents archival evidence of procurement malfeasance. In this figure, four projects (out of several that day) received bidders. According to the procurement law, each project should have at least three bidders. However, projects six and seven fall short. Bidding contractors must also provide several documents to qualify for a project. Such documents include their contractor registration, social–security certificate, labor certification, etc. I obtained the universe of over 5000 registered contractors from the Ministry of Works & Housing and the Ministry of Roads & Highways. In practice, very few winning contractors were actually registered. In Figure 1, contractor SCL possesses all required documents for project five, but is missing key documents for project seven, which would have been submitted on the same day. This is rather suspicious. Junior bureaucrats (who did not benefit from procurement malfeasance), reported that it was common for projects with multiple bidders to have two bidders fail to provide all documents, essentially leaving only one choice. Why design and facilitate such a process when it would be easier to accept a kickback without record? This system ensures all parties involved have plausible deniability: they went through the process, several contractors were clearly not qualified and the bureaucrats can vouch for the quality of the winners. However, just because laws are not being broken, does not mean they are being followed. Party chairs are the enforcers of this system, possessing leverage over politicians, contractors and bureaucrats. From politicians and contractors, the party chair expects kickbacks. In return, party chairs can provide access to contracts and can ensure the politician of the support of the party. With bureaucrats, party chairs expect that they will apply their imprimatur to the procurement process. Otherwise, as mentioned by the DCD above, party sympathizers can “frustrate” the bureaucrat.

“You see, it was the party that got you into power, whether you are the Presi-

13 14

Figure 1: Data on project bidders and their qualiﬁcations from a southern district. dent or an MP. In the Party, we have what are called ‘foot soldiers’. They go into the very rural, remote areas, and they canvas for votes. They go places where President and MPs cannot go. That is why they are important, and that is why Accra listens to them.” District Chief Executive, Forest Zone

Figure 2 summarizes these actions visually as an “iron table”. Contractors exchange promised kickbacks for access to contracts, which they receive from politicians, bureaucrats and party chairs. For bureaucrats, they provide their expert legitimacy, which ensures that politicians and chairs do not interfere with their job security. Bureaucrats, motivated by retirement and social pressures, value that job security and the income provided by contractors. Politicians provide income to the party chair, and in turn the chair ensures that the politician can have electoral ﬁnance and the support of party foot soldiers.

C. Enforcement

How is this system enforced? What prevents a politician or contractor from refusing to pay their kickbacks? The party chairs play a critical role in enforcing this equilibrium.

“With the MPs, our relationship is good. I helped many of them to win—we assisted by buying motos and giving out cash. I would also talk to the people.” Regional Party Chair, National Democratic Congress

Regional party chairs provide much electoral support, financial and logistical, to incumbents and aspirants.21 Chairs mobilize the party’s foot soldiers. In turn, it is expected that politicians will pay rents to these chairs when they win office. These rents allow chairs to pay foot soldiers and give them jobs (often in unskilled construction labor), similar to Tammany Hall’s machine. Critically, if a politician fails to pay rents, the chair can mobilize foot soldiers against the politician and support another willing candidate for the office.

“The Regional Chair mobilized youth and they vandalized the Regional Coor- dinating Council. They destroyed the tender box. The Regional Chair has all the projects in this region, but only one project is executed by himself. They give these out on percentage basis. They become a party chairman to enrich themselves.” Constituency Party Secretary, National Democratic Congress

21Many aspirants also take out bank loans to help ﬁnance their campaigns. Paying these loans also leads them to extract contract kickbacks once in oﬃce.

Access to Contracts Income Chair

Contractor Income

Election Finance, Mobilization Income Job Security

Expert LegitimacyExpert Access to Contracts

Income Access to Contracts to Access

Expert Legitimacy ce o C tacs, ob Security b Jo , cts ntra Co to s Acces Politician Bureaucrat The Iron Table Model of Political Organization. Contracts are ﬁnanced by the central government and Figure 2: through pooled external donor funds (DDF).

16 MP DCE MP DCE Chair RM DCE MP DCE Money Flow Supporting Foot Soldiers Supporting Attacking Foot Soldiers A visual model of the relationship between a regional party chair and the politicians from a given region. Figure 3: Politicians provide income toblue) who the support chair the fromof politicians’ sending their ambitions. foot own If soldiers a contractdisarm to politician kickbacks, the destabilize defects and chair’s the from credible the politician’s this threat. constituency game, chair (in the red). provides chair Politicians foot holds lack soldiers the incentive (in credible to threat coordinate to

17 Figure 3 presents this interaction between a regional chair and politicians. In a given region, a party chair oversees multiple politicians. Each chair controls the region’s foot soldiers who mobilize in support of a party’s candidates. In exchange, foot soldiers expect chairs to provide money and jobs. Chairs finance this through the kickbacks received from the various politicians in their region, and chairs often have multiple construction companies of their own. Party chairs know when projects arrive from Accra, and they compel each politician to award projects to favored contractors—both the chair and politician receive kickbacks from this event. A region’s politicians are locked in a multiplayer Prisoners’ Dilemma with each other: it would take a critical mass of them to stop funding a chair to disable the credible threat. However, it is in each politician’s individual interest to kick funds back to the chair to ensure that foot soldiers are not used against them. Figure 3 shows this coordination failure. The Regional Minister, a DCE and an MP failed to supply enough money, and foot soldiers were sent to destabilize those offices (in red). So a politician faces these party pressures, but it is not the politician alone who awards the contracts. By the design of the procurement law, bureaucrats must still apply their expert imprimatur.22 Bureaucrats are motivated by a desire to finance retirement, but they also value job security and stability.

“The pressure transcends from DCE to staﬀ. You transfer it to staﬀ in ensuring they do what is right. [...] I want to favor you, this is the process and we must act contrary to what is right.” Member of Parliament, National Democratic Congress

In a functioning democracy, a bureaucrat could turn into a whistleblower, reporting illicit practices to superiors in the capital. This is not the case in Ghana. The central bureaucracy in the capital has been politicized.

“As internal auditor, I know it. And external auditors, they know it. With me, I cannot get my way, because my report goes to DCE. Even if I send report to internal—audit agency, he (DCE) knows someone there—and he can call them to change it.” Internal Auditor, District G

“One of the areas that give us so much trouble. Everyone wants contracts from us. [...] Most times (80%), you don’t have opportunity to defend yourself.

22Indeed, having a thorough procurement law is seen favorably by donors.

18 Health Budget Planner Finance Engineer Education Agriculture Players Set Coalition Minimum Yielding DCD DCE Clinic Dorm Projects A visual model of the coalition a DCE can build with her bureaucrats to divide a district’s contracts amongst Public Public WC Feeder Road Market Stalls School Block Figure 4: themselves. Including a few of the more powerful bureaucrats keeps the remaining bureaucrats in line.

19 If you are here, they will bring a letter and you are transferred.” District Coordinating Director, District K

The threat of transfer strongly compels bureaucrats to fall into line. However, it is imprac- tical for a politician, namely a DCE, to threaten the transfer of every bureaucrat around her. She must share some of projects with at least a few critical bureaucrats to ensure that all bureaucrats under her yield. Essentially, the DCE is playing an ultimatum game with her bureaucrats to form a minimum winning coalition that ensures stability.

“Most of corruption is from the civil servants. They are always in the system— they have to get something out of the project. Normally, they want someone they can manipulate to get something out of it.” Former DCE, New Patriotic Party

Figure 4 depicts this game. This hypothetical district is awarding six projects. Ideally, a DCE would prefer to keep all six projects for her preferred contractors, but this might actuate retaliation from the bureaucrats. Instead, the DCE keeps two projects for herself, and awards two to the District Coordinating Director, who is the most senior bureaucrat. The DCD can more credibly ensure that the other senior bureaucrats yield. Oﬃcers report that it can be diﬃcult to judge how a DCE will distribute projects.

“I just came to this district—you have to study terrain before you work. You have to be very careful.” District Finance Oﬃcer, District E

“See, this system is highly monopolized. DCD and DFO monopolize the district assembly. These two people can sabotage—when DCE joins them it is worse. Other times, DCE and DFO. Depends on the district: two people are hijacking projects. High costs from DWE inﬂate prices. Any district—DPO is always working—some are chasing money.” District Planning Oﬃcer, District E

In this example, the Engineer and Finance Officer receive a project each for their preferred contractors. In most districts of my sample, these two senior officers were reported most frequently by others as the ones profiting from procurement. This may relate to their specializations. The Engineer oversees procurement, and would be in contact with any contractors that other bureaucrats favor. The Finance Officer oversees payment, not just for procurement, but also for per diems when bureaucrats travel for work. The other senior

20 bureaucrats have little incentive to protest against these two officers. Depending on the number of projects awarded, other officers, such as the planner and budget officer, may receive a project. As for the remaining officers, who receive zero projects, they are not any worse off and still hold their jobs, so they should yield to this equilibrium.

D. Continuity

The actions specified above are performed in a context where contracts are flowing from the central government. But Ghana has two viable parties that have both been in government. How does the other party stay financed when it is not being awarded public contracts? Arriola [2012] theorizes that the private sector will be more likely to finance opposition coalitions in an environment of financial liberalization. I theorize that the opposition in Ghana stays financed because of contractors’ economic fears. A contractor fears that a change in government will lead to a stoppage in the supply of contracts. As a result, a contractor will subcontract portions of the work to opposition contractors to ensure that these contractors remember him when they are in power. These friends will supply him with work should there be a change in party government. Effectively, contractors are buying insurance.

“We are like one family, all one tribe. If I have a job and don’t have equipment, you can help—cement, roller, etc. When my cement comes, I release it back to you. It is cordial. Sometimes we’ll subcontract, especially with electricals. NDC guys help out NPP now for help in the future.” Contractor, District C

It has also been reported that a party foot soldier will win a contract, when in fact that foot soldier is not a contractor at all. That foot soldier will then auction oﬀ the contract (after taking his monetary share), and often an opposition contractor will buy it.

“With the subletting (subcontracting), some of the (NDC) party activists, we call them ‘foot soldiers’, win, but they have no certiﬁcates or skills. What happens when contract is given to them? They sublet it, often to the NPP contractor. Why? Well, the NDC contractors say, ‘I’m an NDC man, but I may not have good relations with the procurement unit of the Regional Minister.’ But in subletting they might need money—that is underneath, and the public does not see it. The foot soldier can sell the contract to an NDC or NPP man

21 to buy. But, sometimes, the NDC man will say, ‘Hey, this is my party, I voted it in. I will not pay the 20,000.”’ Contractor, District D

Contractors also have side businesses to ensure that they stay afloat. Many will sell construction materials, or have particular types of equipment that they can lease to other contractors who do win projects. Many report that contractors switch sides between elections by changing ownership or business names. It is also the case that contractors, once attached to a rising politician, will follow that politician as she ascends to higher office, effectively forming a political action committee. Ghana’s constitution also ensures that both major parties stay viable across the country. Under the constitution, the President chooses the DCE for every district. Essentially, the NDC can extract contract rents in the heart of an NPP stronghold, and vice versa. Thus, contractors who are loyal to a minority party in a given area, can still extract valuable insurance payments from their friends on the other side. These contractors can use that money to win the favor of politicians from their own party once they come into power.

IV. Implications

In Ghana, politicians and bureaucrats monitor the building of local development projects. When they are compromised by kickbacks, the quality of development suﬀers.

“It is because the Assembly oﬃcers have an interest in it. Say this building we are in, the contract sum is 10,000. The contractors will not get 10,000. Here, we moved in January, two months ago! Shoddy work! Look at that door! Even the director does not have a door. Is there a place of convenience here? No!” District Education Oﬃcer, District E

Portions of the corruption mechanism described above are “open secrets” in Ghana. One would presume that social audits, where the population shames errant politicians and bureaucrats, might lessen the prevalence of this behavior. Unfortunately, one NDC MP described Ghana to me as having a “culture of silence”, where citizens were afraid to criticize politicians for fear of “being insulted”. Even traditional chiefs, well respected by the people, are not immune to such verbal attacks.

“Yes, there is shoddy work. The District Assembly awards, but they do not involve the chiefs. Or the contract may be awarded by the region or by Accra.

22 You cannot just tell a contractor that he is doing a bad job. He will report you to the DCE and you will get insulted. [...] On the outside, we chiefs have to welcome all contractors and pledge our support. It is our duty to ensure that all people in our villages are happy and well cared for. But on the inside we are weeping.” Chief, Northern Zone

For chiefs, there is the additional problem that criticizing politicians and contractors today, might result in few or no projects in your village tomorrow. It is better to have a half–built school than no school at all. Plausible deniability is a critical component of this equilibrium, and it manifests itself in another way. In work I present elsewhere, bureaucrats report that they actually receive the most project requests from local District Assembly members, who represent individual villages. In my participant observations, I found that bureaucrats generally performed a thorough job vetting these requests and identifying areas of most need. Certainly, DCEs and MPs steered occasional projects to their supporters’ areas, but I contend that that is not abnormal behavior given that citizens elected these politicians for (amongst others) development reasons. However, the corrupt mechanism I identify carries a sinister im- plication: procurement corruption turns targetable, physical projects into liquid electoral capital that can be used anywhere in any fashion. In theory, it is possible for projects to be targeted optimally to the neediest populations and for bureaucrats to believe that they are serving the neediest. They could plausibly deny that they were ignoring the most impoverished or that they were only satisfying one party’s constituents. But targeting a project is only part of the story. Who builds the project is just as important as what the project is. A politician can claim credit for building these projects—which stand just long enough to influence an election—and then use the money from kickbacks to buy votes and silence opposition. For development to serve the needs of the people, it must first serve the interests of the political masters. Ghana is often viewed as an example of democratization for other developing countries. However, as elections become more competitive, the demand for political finance increases. In future work, I will analyze whether such increased politician and party demand for finance results in increased procurement corruption and lower quality of development. Numerous bureaucrats reported that this procurement malfeasance did not exist more than 10 years ago, namely because officials feared former military dictator–

23 turned–President Jerry Rawlings.23 Future work will examine alternative modes of political finance and bureaucratic–retirement plans to determine whether meaningful alteration in behavior is possible. For bureaucrats, is there a way to connect retirement savings with long–term performance and credible monitoring of projects? A solution to corruption does not necessarily mean eradicating corruption entirely—such a result is likely unachievable in the developed context. Even if a politician selects a favored contractor, can a system be designed to ensure that the favorite performs satisfactorily?24 Future work will build on this theory by comparing it to evidence from other countries. The portable questions to be asked: how is politics financed in other developing contexts? What is the nature of the procurement process? In Ghana, procurement provides an avenue through which politics can be financed in a way that is difficult to prosecute. In other countries, variations in political finance can impose different effects on the quality of public goods and services.

V. Conclusion

Political finance defines political organization; political organization shapes development. In Ghana, four actors—politicians, bureaucrats, contractors, party chairs—collude to extract rents from the public–procurement process to satisfy their personal ambitions, notably the financing of political parties and campaigns. The bureaucrat ensures that the procurement process is followed on paper (though in reality collusion, cyclical bidding, etc., occur), and her monopoly on expert discretion obstructs prosecution. The political–party chair applies pressure to politicians for rents because he holds a credible threat to mobilize party supporters against recalcitrant politicians. Politicians then transfer such pressure onto the bureaucrats, with whom they form minimum coalitions to ensure an optimal allocation of contracts. Ultimately, the quality of development suffers as these key actors lack incentive to monitor projects. As elections become more competitive, the demand for political finance will increase, which, if still financed from contract kickbacks, will only ensure further suboptimal development—a vicious cycle. In the last two years, discontent in Ghana has grown. Unfortunately, democratization is a reflection of human nature, containing both the path to better outcomes and the seeds of its own demise.

23For an alternate perspective on Ghana’s party development, see Riedl [2014]. 24In the American context, many large infrastructure projects were built with favored immigrant labor, but those projects still stand today (cf. Erie 1992). I am still investigating why that is the case.

24 VI. References

Achebe, Chinua. 1960. No Longer at Ease. (London: Heinemann).

Aldrich, John. 1995. Why Parties? The Origin and Transformation of Political Par- ties in America. (Chicago: University of Chicago Press).

Arriola, Leonard. 2012. Multi–Ethnic Coalitions in Africa: Business Financing of Op- position Election Campaigns. (New York: Cambridge University Press).

Bates, Margaret Rouse. 1971. UNIP in Post–Independence Zambia: The Development of an Organizational Role. Harvard University, Department of Government. Unpublished PhD Dissertation.

Bawn, Kathleen, et al. 2012. “A Theory of Political Parties: Groups, Policy Demands and Nominations in American Politics.” Perspectives on Politics, 10(3).

Blundo, Giorgio & Jean–Pierre Olivier de Sardan. 2006. Everyday Corruption and the State. (London: Zed Books).

Carpenter, Daniel. 2001. The Forging of Bureaucratic Autonomy: Reputations, Networks and Policy Innovation in Executive Agencies, 1862-1928. (Princeton, NJ: Princeton Uni- versity Press).

Erie, Stephen. 1992. Rainbow’s End: Irish Americans and the Dilemmas of Urban Ma- chine Politics. (Berkeley: University of California Press).

Fenno, Richard. 1978. Home Style: House Members in Their Districts. (New York, NY: Addison–Wesley).

Fiorina, Morris. 1977. “The Case of the Vanishing Marginals: The Bureaucracy Did It.” American Political Science Review, 71(2).

Fiorina, Morris & Roger Noll. 1978. “Voters, Bureaucrats and Legislators: A Rational–

25 Choice Perspective on the Growth of Bureaucracy.” Journal of Public Economics, 9.

Friederich Ebert Stiftung. 2010. “A Guide to District Assemblies in Ghana.”

Glynn, Adam. 2013. “What Can We Do With Statistical Truth Serum? Design and Analysis of the List Experiment.”

Hainmueller, Jens, et al. 2013. “Causal Inference in Conjoint Analysis: Understanding Multidimensional Choices via Stated Preference Experiments.”

Heberlig, Eric & Bruce Larson. 2012. Congressional Parties, Institutional Ambition and the Financing of Majority Control. (Ann Arbor: University of Michigan Press).

Hodgkin, Thomas. 1961. African Political Parties. (Baltimore: Penguin Books).

Huntington, Samuel. 1968. Political Order in Changing Societies. (New Haven: Yale University Press).

Kramon, Eric. 2011. “Why Do Politicians Buy Votes When the Ballot Is Secret? Theory and Experimental Evidence from Kenya.”

Lessig, Lawrence. 2011. Republic, Lost: How Money Corrupts Congress—And a Plan to Stop It. (New York, NY: Hachette).

Lindberg, Staﬀan. 2010. “What Accountability Pressures Do MPs in Africa Face and How Do They Respond? Evidence from Ghana.” Journal of Modern African Studies, 48(1).

Lipsky, Michael. 1989. Street-Level Bureaucracy: Dilemmas of the Individual in Public Service. (New York, NY: Russell Sage Foundation).

Mayhew, David. 1974. Congress: The Electoral Connection. (New Haven, CT: Yale University Press).

26 Michels, Robert. 1915. Political Parties: A Sociological Study of the Oligarchical Ten- dencies of Modern Democracy. (London: Jarrold & Sons).

Mudge, Stephanie and Anthony Chen. 2014. “Political Parties and the Sociological Imag- ination: Past, Present and Future Directions.” Annual Review of Sociology, 40.

Ostrogorski, Moishe. 1895. Democracy and the Organization of Political Parties: Eng- land. (Piscataway, NJ: Transaction).

Riedl, Rachel Beatty. 2014. Authoritarian Origins of Democratic Party Systems in Africa. (New York: Cambridge University Press).

Schlesinger, Joseph. 1966. Ambition and Politics: Political Careers in the United States. (Chicago, IL: Rand McNally).

Weber, Max. 1919. “Politics as a Vocation.”

Weingast, Barry. 1979. “A Rational–Choice Perspective on Congressional Norms.” Amer- ican Journal of Political Science, 23(2).